U.S. patent application number 09/814622 was filed with the patent office on 2001-09-06 for method and apparatus for analyzing data and advertising optimization.
Invention is credited to Cannon, Mark E..
Application Number | 20010020236 09/814622 |
Document ID | / |
Family ID | 21899617 |
Filed Date | 2001-09-06 |
United States Patent
Application |
20010020236 |
Kind Code |
A1 |
Cannon, Mark E. |
September 6, 2001 |
Method and apparatus for analyzing data and advertising
optimization
Abstract
The most preferred embodiment of the present invention is a
computer-based decision support system that includes three main
components: a database mining engine (DME); an advertising
optimization mechanism; and a customized user interface that
provides access to the various features of the invention. The user
interface, in conjunction with the DME, provides a unique and
innovative way to store, retrieve and manipulate data from existing
databases containing media-related audience access data, which
describe the access habits and preferences of the media audience.
By using a database with a simplified storage and retrieval
protocol, the data contained therein can be effectively manipulated
in real time. This means that previously complex and lengthy
information retrieval and analysis activities can be accomplished
in very short periods of time (typically seconds instead of minutes
or even hours). Further, by utilizing the advertising optimization
mechanism of the present invention, businesses, networks, and
advertising agencies can interactively create, score, rank and
compare various proposed or actual advertising strategies in a
simple and efficient manner. This allows the decision-makers to
more effectively tailor their marketing efforts and successfully
reach the desired target market while conserving scarce advertising
capital. Finally, the user interface for the system provides access
to both the DME and the optimization mechanism in a simple and
straightforward manner, significantly reducing training time.
Inventors: |
Cannon, Mark E.; (Provo,
UT) |
Correspondence
Address: |
Schmeiser, Olsen & Watts LLP
18 East University Drive, #101
Mesa
AZ
85201
US
|
Family ID: |
21899617 |
Appl. No.: |
09/814622 |
Filed: |
March 22, 2001 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09814622 |
Mar 22, 2001 |
|
|
|
09038380 |
Mar 11, 1998 |
|
|
|
Current U.S.
Class: |
1/1 ;
707/999.001 |
Current CPC
Class: |
H04H 60/66 20130101;
H04N 21/812 20130101; H04H 60/45 20130101; G06Q 30/0201 20130101;
H04N 21/25883 20130101; H04N 21/478 20130101; G06Q 30/02 20130101;
H04H 60/46 20130101; H04N 21/44222 20130101; G06Q 30/0269
20130101 |
Class at
Publication: |
707/1 |
International
Class: |
G06F 007/00 |
Claims
1. An apparatus comprising: a CPU; a memory coupled to the CPU; an
advertising optimization mechanism residing in the memory and being
executed by the CPU, the advertising optimization mechanism
iteratively modifying and scoring a base advertising schedule in
order to achieve an optimal advertising schedule.
2. The apparatus of claim 1 further comprising a graphical user
interface with a plurality of icons which provide a plurality of
choices for advertising optimization.
3. The system of claim 1 further comprising at least one index
residing in the memory and cooperating with the the advertising
optimization mechanism to iteratively modify and score the base
advertising schedule.
4. The apparatus of claim 3 wherein the at least one index
comprises at least one of an exposure valuation index, an audience
valuation index, an exposure recency index, a response index and a
cost index.
5. The apparatus of claim 1 further comprising a database mining
engine residing in the memory.
6. The apparatus of claim 5 wherein the database mining engine
further comprises a plurality of Boolean filters used to screen the
plurality of person-by-person records contained in the
database.
7. The apparatus of claim 1 further comprising a data conversion
mechanism residing in the memory.
8. The apparatus of claim 7 wherein the data conversion mechanism
comprises a mechanism to convert data from a first data format to a
second data format.
9. The apparatus of claim 8 wherein the first data format is a
plurality of television viewing records received from A. C. Nielsen
and the second data format is a binary representation of the
plurality of television viewing records.
10. The system of claim 1 further comprising a plurality of indices
residing in the memory and cooperating with the the advertising
optimization mechanism to iteratively modify and score the base
advertising schedule.
11. The apparatus of claim 3 wherein the plurality of indices
comprises at least two of an exposure valuation index, an audience
valuation index, an exposure recency index, a response index and a
cost index.
12. A computer system for optimizing an advertising schedule, the
computer system comprising: a CPU; a memory coupled to the CPU; a
database residing in the memory, the database containing a
plurality of person-by-person data files, the plurality of
person-by-person data; a database mining engine residing in the
memory; a data conversion mechanism residing in the memory, the
data conversion mechanism comprising a mechanism for converting
data from a first data format to a second data format; and a
graphical user interface residing in the memory and being executed
by the CPU, wherein the graphical user interface provides a
plurality of choices for optimizing the advertising schedule
according to a plurality of indices.
13. The computer system of claim 12 wherein the first data format
is a plurality of television viewing records received from A. C.
Nielsen and the second data format is a binary representation of
the plurality of television viewing records.
14. The computer system of claim 12 wherein the plurality of
indices includes an exposure valuation index, an audience valuation
index, an exposure recency index, a response index and a cost
index.
15. A program product comprising: an advertising optimization
mechanism, the advertising optimization mechanism iteratively
modifying a base advertising schedule to achieve an optimal
advertising schedule; and signal bearing media bearing the
advertising optimization mechanism.
16. The program product of claim 16 wherein the signal bearing
media comprises transmission media.
17. The program product of claim 16 wherein the signal bearing
media comprises recordable media.
18. The program product of claim 16 further comprising a plurality
of indices which are utilized by the advertising optimization
mechanism to iteratively modify the base advertising schedule.
19. The program product of claim 18 wherein the plurality of
indices comprises an exposure valuation index, an audience
valuation index, an exposure recency index, a response index and a
cost index.
20. The program product of claim 15 further comprising a data
conversion mechanism, the data conversion mechanism comprising a
mechanism for converting data from a first data format to a second
data format.
21. The program product of claim 20 wherein the first data format
is a plurality of television viewing records received from A. C.
Nielsen and the second data format is a plurality of variable
length records which describe changes in media-related access data
for a target audience.
22. The program product of claim 20 wherein the first data format
is a plurality of television viewing records received from A. C.
Nielsen and the second data format is a binary representation of
the plurality of television viewing records.
23. A method for advertising optimization, the method comprising
the step of iteratively modifying a base advertising schedule
according to at least one of a plurality of indices in order to
achieve an optimal advertising schedule.
24. The method of claim 23 wherein the plurality of indices
comprises an exposure valuation index, an audience valuation index,
an exposure recency index, a response index and a cost index.
25. The method of claim 23 wherein the step of iteratively
modifying a base advertising schedule comprises using a weighted
effective frequency method to score and compare a plurality of
possible alternative advertising schedules.
26. The method of claim 25 wherein the step of scoring and
comparing a plurality of possible alternative advertising schedules
comprises the step of assigning a value to a modified advertising
campaign based on previous or anticipated individual or collective
advertising exposure.
27. The method of claim 23 wherein the step of iteratively
modifying a base advertising schedule comprises using a time
weighted effective frequency method to score and compare a
plurality of possible alternative advertising schedules.
28. A computer-implemented method, the method comprising the steps
of: (a) providing an advertising campaign containing a plurality of
advertising spots; (b) identifying one of the plurality of
advertising spots as a least valuable advertising spot; (c)
removing the least valuable advertising spot from the advertising
campaign; (d) identifying a plurality of alternative options to add
to the advertising campaign; (e) selecting one of the plurality of
alternative options and adding the selected alternative option to
the advertising campaign to achieve a modified advertising
campaign; (f) scoring the modified advertising campaign; and (g)
repeating steps b, c, d, e, and f in order to achieve an optimal
advertising schedule.
29. The method of claim 28 wherein the step of scoring the modified
advertising campaign comprises the step of using a weighted
effective frequency method to score the modified advertising
campaign.
30. The method of claim 28 wherein the step of scoring the modified
advertising campaign comprises the step of using a time weighted
effective frequency method to score the modified advertising
campaign.
31. The method of claim 28 wherein the step of scoring the modified
advertising campaign comprises the step of using at least one index
to score the modified advertising campaign.
32. The method of claim 31 wherein the step of scoring the modified
advertising campaign using at least one index to score the modified
advertising campaign comprises the step of using a plurality of
indices to score the modified advertising campaign.
33. The method of claim 32 wherein the step of scoring the modified
advertising campaign using a plurality of indices comprises the
step of using at least two of an exposure valuation index, an
audience valuation index, an exposure recency index, a response
index and a cost index to score the modified advertising
campaign.
34. The method of claim 33 further comprising the step of using a
series of product usage data as an input for the response
index.
35. A graphical user interface comprising at least one icon which
accesses a plurality of person-by-person records contained in a
database via a database mining engine and presents at least one
advertising optimization choice to a user of the graphical user
interface.
36. The graphical user interface of claim 35 further comprising a
scoring mechanism which provides a score for an advertising
campaign based on a plurality of indices.
37. The graphical user interface of claim 36 wherein the scoring
mechanism uses a plurality of indices to score the advertising
campaign.
38. The graphical user interface of claim 37 wherein the plurality
of indices comprises at least two of an exposure valuation index,
an audience valuation index, an exposure recency index, a response
index and a cost index.
39. A computer system with a graphical user interface comprising: a
CPU; a memory coupled to the CPU; a database residing in the
memory, the database comprising a plurality of person-by-person
media-related records which describe a series of choices and
decisions made by an identified sample audience in relation to a
media vehicle; a database mining engine residing in the memory and
being executed by the CPU; and at least one icon which accesses the
plurality of person-by-person records contained in the database via
the database mining engine and presents at least one advertising
optimization choice to a user of the graphical user interface.
40. The computer system of claim 39 further comprising a scoring
mechanism residing in the memory, the scoring mechanism providing a
score for an advertising campaign.
41. A computer system for analyzing data and optimizing an
advertising schedule, the system comprising: a CPU; a memory
coupled to the CPU; a database residing in the memory, the database
comprising a plurality of person-by-person records which describe a
series of television choices and decisions made by an identified
sample audience; a database mining engine residing in the memory,
the database mining engine comprising a plurality of Boolean
filters used to screen the plurality of person-by-person records
contained in the database; and a graphical user interface residing
in the memory and being executed by the CPU, wherein the user
interface accesses the person-by-person records in the database via
the database mining engine and iteratively optimizes the
advertising schedule using a predetermined method.
42. The computer system of claim 41 wherein the predetermined
method is a weighted effective frequency method.
43. The computer system of claim 41 wherein the predetermined
method is a time weighted effective frequency method.
44. The computer system of claim 41 further comprising a data
conversion mechanism, the data conversion mechanism comprising a
mechanism for converting data from a first data format to a second
data format.
45. The computer system of claim 44 wherein the first data format
is a plurality of television viewing records received from A. C.
Nielsen and the second data format is a binary representation of
the plurality of television viewing records.
46. A method of calculating a ratio, the method comprising the
steps of: generating a first media-related exposure value;
generating a second media-related exposure value; and combining the
first and second media-related exposure values to create the
ratio.
47. The method of claim 46 wherein the step of combining the first
and second media-related exposure values to create the media
analysis ratio comprises the step of dividing the first
media-related exposure value by the second media-related exposure
value.
48. The method of claim 46 wherein the step of generating the first
media-related exposure value comprises the step of selecting a
subset of person-by-person media-related access data from a
database.
49. The method of claim 46 wherein the step of generating the
second media-related exposure value comprises the step of selecting
a subset of person-by-person media-related access data from a
database.
50. A method of calculating a media analysis ratio, the method
comprising the steps of: selecting a subset of person-by-person
media-related access data from a database thereby generating a
first media-related exposure value; selecting a subset of
person-by-person media-related access data from the database
thereby generating a second media-related exposure value; and
dividing the first media-related exposure value by the second
media-related exposure value to create the media analysis
ratio.
51. A method of scoring an advertisement, the method comprising the
steps of: scoring each of a plurality of individual exposures to
the advertisement to determine a value for each of the plurality of
individual exposures; and combining the values determined for each
of the plurality of individual exposures to achieve an overall
score for the advertisement.
52. The method of claim 51 wherein the step of scoring each of the
plurality of individual exposures to an advertisement to determine
a value for each of the plurality of individual advertising
exposures comprises the step of using a plurality of factors in
combination to score each of the plurality of individual exposures
to an advertisement.
53. A computer system with a graphical user interface comprising: a
CPU; a memory coupled to the CPU; a database residing in the
memory, the database comprising a plurality of person-by-person
media-related records which describe a series of choices and
decisions made by an identified sample audience in relation to a
media vehicle; a database mining engine residing in the memory and
being executed by the CPU; and a graphical user interface with at
least one icon which accesses the plurality of person-by-person
records contained in the database via the database mining engine
and presents at least one advertising optimization choice to a user
of the graphical user interface.
54. The computer system of claim 53 wherein the graphical user
interface further comprises a mechanism for evaluating a plurality
of alternative advertising options.
55. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for distributing advertisements over time and space
based on actual or anticipated individual or collective advertising
exposure.
56. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for assigning advertising response values to a
plurality of media alternatives.
57. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for assigning costs to the plurality of alternative
advertising options based on time or space boundaries for the
purpose of scoring the plurality of alternative advertising
options.
58. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for assigning individual exposure values to the
plurality of alternative advertising options according to the value
of at least one of a plurality of individual demographic
measurements.
59. The computer system of claim 58 wherein the mechanism for
assigning individual exposure values comprises a mechanism for
displaying the individual exposure values of the at least one of a
plurality of individual demographic measurements.
60. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for displaying the estimated influence of advertising
messages based on the declining influence of advertising over
time.
61. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for displaying the estimated influence of advertising
messages based accumulated advertising messages over time.
62. The computer system of claim 54 wherein the mechanism for
evaluating a plurality of alternative advertising options comprises
a mechanism for assigning advertising value to multiple levels of
advertising exposure based on frequency of exposure.
63. The computer system of claim 62 wherein the mechanism for
assigning advertising value to multiple levels of advertising
exposure based on frequency of exposure further comprises a
mechanism for displaying the assigned advertising values.
64. The computer system of claim 62 wherein the mechanism for
assigning advertising value to multiple levels of advertising
exposure based on frequency of exposure comprises a mechanism for
assigning advertising value to multiple levels of advertising
exposure based on actual or anticipated exposure to an
advertisement.
65. A method for comparatively scoring a plurality of advertising
options comprising the step of using a graphical user interface to
evaluate a plurality of alternative advertising options.
66. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of distributing advertisements over time
and space based on actual or anticipated individual or collective
advertising exposure.
67. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of assigning advertising response values
to a plurality of media alternatives.
68. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of assigning costs to the plurality of
alternative advertising options based on time or space boundaries
to score each of the plurality of alternative advertising
options.
69. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of assigning individual exposure values
to each of the plurality of alternative advertising options
according to the value of at least one of a plurality of individual
demographic measurements.
70. The computer system of claim 69 wherein the step of assigning
individual exposure values to each of the plurality of alternative
advertising options according to the value of at least one of a
plurality of individual demographic measurements comprises the step
of displaying the individual exposure values of the at least one of
a plurality of individual demographic measurements.
71. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of displaying the estimated influence of
advertising messages based on the declining influence of
advertising over time.
72. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of displaying the estimated influence of
advertising messages based accumulated advertising messages over
time.
73. The method of claim 65 wherein the step of using a graphical
user interface to evaluate a plurality of alternative advertising
options comprises the step of assigning advertising value to
multiple levels of advertising exposure based on frequency of
exposure.
74. The method of claim 73 wherein the step of assigning
advertising value to multiple levels of advertising exposure based
on frequency of exposure comprises the step of displaying the
assigned advertising values.
75. The method of claim 73 wherein the step of assigning
advertising value to multiple levels of advertising exposure based
on frequency of exposure comprises the step of assigning
advertising value to multiple levels of advertising exposure based
on actual or anticipated exposure to an advertisement.
76. A method of calculating a score for an advertising spot, the
method comprising the steps of: determining a separate value for
each exposure of each of a plurality of audience members to the
advertising spot; and summing the exposure values for each of the
plurality of audience members to calculate the score for the
advertising spot.
77. The method of claim 76 wherein the step of determining a value
for each exposure of each of a plurality of audience members to the
advertising spot comprises the step of a using a weighted effective
frequency method to determine a value for exposing each of a
plurality of audience members to the advertising spot.
78. The method of claim 76 wherein the step of determining a value
for each exposure of each of a plurality of audience members to the
advertising spot comprises the step of a using a time weighted
effective frequency method to determine a value for exposing each
of a plurality of audience members to the advertising spot.
79. The method of claim 76 wherein the step of determining a value
for each exposure of each of a plurality of audience members to the
advertising spot comprises the step of a using predetermined
formula to determine a value for each exposure of each of a
plurality of audience members to the advertising spot.
80. The method of claim 79 wherein the step of a using
predetermined formula to determine a value for each exposure of
each of a plurality of audience members to the advertising spot
comprises the step of using the formula 32 S b ( a ) = i = 1 N a [
V I n ( i ) .times. d = 1 D V A d ( i ) ] .times. V T ( a ) .times.
V R ( a ) V C ( a ) to determine a value for each exposure of each
of a plurality of audience members to the advertising spot.
81. The method of claim 76 wherein the step of summing the exposure
values for each of the plurality of audience members to calculate
the score for the advertising spot comprises the step of using a
using predetermined formula to sum the exposure values for each of
the plurality of audience members.
82. The method of claim 81 wherein the step of the step of using a
using predetermined formula to sum the exposure values for each of
the plurality of audience members comprises the step of using the
formula 33 i = 1 N a [ V I n ( i ) .times. d = 1 D V A d ( i ) ] to
sum the exposure values for each of the plurality of audience
members.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] The present invention relates to the field of
computer-assisted data manipulation and analysis. More
particularly, the present invention relates to methods and
techniques for quickly and efficiently accessing and sorting large
quantities of demographic data and media access information for
various decision-making purposes, especially advertising.
[0003] 2. Description of Related Art
[0004] The business of supplying information and/or entertainment
mingled with paid advertising to audiences has become an important
part of the world economy. Many large corporations have as their
primary business to inform or entertain their customers through
media. In addition, most businesses depend on advertising to reach
potential customers with product or service information. Television
viewing, for example, is one of the most popular activities in
homes around the world. Large numbers of people both in the United
States and abroad use the television as a primary source for news,
education, entertainment, and various social activities. This large
population of television viewers is also a very desirable group of
consumers, sought after for their purchasing power. Most businesses
and companies recognize the potent power of the television when it
comes to attracting and retaining consumers for their various
products and services. Television has the capability to transmit
virtually any message to millions of people in an instant. Because
of this enormous potential, television advertising is believed to
be one of the most important advertising vehicles available for
reaching a desired consumer population.
[0005] Because television has an enormous potential to reach and
influence many consumers, many businesses spend large amounts of
money on television advertising, thereby making advertising revenue
a major source of income for both the networks which create and
commission television programs and the television stations which
broadcast them. Broadcast television advertising revenue is
estimated to be in excess of $30 billion per year, which explains
why advertisers care so much about television viewing. The most
important part of the equation is to try and get the advertising
message in front of the right group of television-viewing
consumers, i.e., matching the product with the desired target
market.
[0006] Since the value of target marketing is well known, the focus
of most businesses is to try and place their advertisements in
commercial slots on television shows that effectively attract and
retain the targeted consumer groups. For example, many major
sporting events are heavily subsidized by advertising campaigns and
commercials promoting beer. On the other hand, most weekly home and
garden programs don't present any advertising or commercial
promotions for alcoholic beverages. Obviously, the businesses that
advertise and market alcoholic beverages such as beer have
determined that the consumers who purchase beer are more likely to
be watching Monday Night Football than Martha Stewart's Living. By
focusing beer-related advertising efforts on those programs that
the target audience is most likely to watch, the advertising
campaign will, in theory, attract more consumers and pay greater
dividends.
[0007] To effectively determine which shows are most favored by the
desired target market, advertising agencies and businesses have
utilized the services of various different research and consulting
firms. These firms purportedly have the ability to accurately
identify which segment of the consumer population is most likely to
be viewing which television program at any given time. In addition,
these research firms try to predict which viewers will be most
receptive to various advertising campaigns, based on the
demographic make-up of the viewing population. Based upon the
weekly viewing information prepared and presented by the television
viewing-related research firms/agencies, advertising campaigns are
born and terminated. Above all, however, these advertising
campaigns are most often the result of educated estimates,
well-thought out probabilities, and other experience-based
decision-making processes. It is most desirable to create an
optimal campaign which effectively utilizes a finite combination of
resources to communicate to the target audience. While this goal is
easy to quantify, it is not so easy to achieve and many advertising
campaigns are simply ineffective.
[0008] The reason why some advertising campaigns are successful and
some are not is really very simple. Although the practice of
identifying target markets and developing advertising campaigns
attract those targeted consumers is a fairly developed practice,
the ability to accurately and efficiently measure which advertising
campaigns are most successful and what changes should be made in an
advertising campaign to increase the overall effectiveness of the
advertising campaign is a far-less developed area of industry.
There is no way to effectively gather feedback for an advertising
campaign and to accurately measure or evaluate the performance of
the advertising effort.
[0009] One of the main problems with the currently used models and
techniques for identifying and implementing the most optimal
advertising campaign for a given product using a given advertising
medium is the lack of effective tools for scoring, evaluating, and
comparing alternative advertising strategies. There is no well
known, acceptable technique or method for evaluating, scoring, and
comparing one advertising plan or schedule and strategy with
another. As further described below, this deficiency reduces the
advertising optimization process to a series of estimates and
educated guesses when determining which campaign, from among a
group of similar campaigns, will be most successful in
accomplishing the desired goals.
[0010] Another problem with developing effective media advertising
campaigns is directly related to the technology limitations of
presently implemented systems. Using the currently-available
systems and methods to manipulate and analyze the huge amounts of
data that are available to decision-makers can take days or even
weeks to accomplish. Frequently, the various systems in use today
will provide data that are no longer relevant by the time the data
are generated. In addition, the lack of sophisticated advertising
optimization tools impose artificial limits on advertising agencies
and media planners that are actively involved in the
decision-making process.
[0011] This is particularly true when trying to create and/or
customize a campaign to reach the target market in the most
cost-effective manner for a given advertiser. There are, at
present, no broad-ranging interactive methods or tools available to
the media planner for optimizing an advertising campaign in real
time. Many media planners have the data available to make strategic
decisions regarding advertising, but the available planning tools
do not allow rapid and easy access to the data in an intuitive,
interactive environment. Specifically, known systems focus only on
"effective reach" and do not allow rapid week-to-week analysis of a
unified sample.
[0012] In addition, present tools rely heavily on estimates and
averages as a means of evaluating the impact of a given advertising
plan or schedule. While good estimates may yield reliable results,
good estimates are more of an exception than a rule. This means
that the planning process is more an art than a science, and many
years of trial-and-error experience are required to effectively
determine or even estimate the probable effectiveness of a given
advertising strategy. Because of these factors, many media planners
are relatively ineffective in preparing valuable advertising
strategies until they have several years of experience.
[0013] Another significant drawback of the systems and methods
presently used to analyze television audience viewing data is
limited access. Many analysis and decision support systems
available today are large, expensive computer systems that many
smaller companies cannot afford to purchase. Given limited access
to necessary resources, many companies are forced to pay
high-priced consultants to analyze the relevant data and to provide
access to the desired information. This further limits the value of
the available data for companies without the financial resources to
engage consultants or purchase expensive equipment.
[0014] Without a more effective system for scoring, comparing and
optimizing advertising campaigns for specific needs, advertising
agencies, networks, businesses, and other interested organizations
will continue to be limited in their efforts to produce effective
advertising campaigns. The result will undoubtedly be less than
optimal use of scarce advertising dollars and an unnecessary loss
in revenue for everyone that relies on the presently available
systems/methods for analysis and decision-making purposes.
DISCLOSURE OF THE INVENTION
[0015] According to a preferred embodiment of the present
invention, a method and apparatus for quickly and easily analyzing
large quantities of computer-based media-related data is disclosed.
The data can be manipulated to evaluate, score and optimize an
advertising campaign by interactively comparing many different
options. The most preferred embodiment of the present invention is
a computer-based decision support system that includes three main
components: a database mining engine (DME); an advertising
optimization mechanism; and a customized user interface that
provides access to the various features associated with the system.
In addition, the various preferred embodiments of the present
invention are available for use with any standard personal
computer, making the system available to a much larger group of
decision-making executives than ever before possible.
[0016] The user interface, in conjunction with the DME, provides a
unique and innovative way to store, retrieve and manipulate data
from existing databases containing media-related access data, which
describe the access habits and preferences of the media audience.
By using a database with a simplified storage and retrieval
protocol, the data contained therein can be effectively manipulated
in real time. This means that previously complex and lengthy
information retrieval and analysis activities can be accomplished
in very short periods of time (typically seconds instead of minutes
or even hours).
[0017] Further, by utilizing the advertising optimization mechanism
of the present invention, businesses, networks, and advertising
agencies can interactively create, score, rank and compare various
proposed or actual advertising strategies in a simple and efficient
manner. This allows the decision-makers to more effectively tailor
their marketing efforts and successfully reach the desired target
market while conserving scarce advertising capital. Finally, the
user interface for the system provides access to both the DME and
the optimization mechanism in a simple and straightforward manner,
significantly reducing training time.
[0018] The foregoing and other features and advantages of the
invention will be apparent from the following more particular
description of preferred embodiments of the invention, as
illustrated in the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] The preferred embodiments of the present invention will
hereinafter be described in conjunction with the appended drawings,
where like designations denote like elements, and:
[0020] FIG. 1 is a block diagram of a preferred embodiment of the
present invention;
[0021] FIG. 2 is a flowchart for converting person-by-person
information from one format to another in accordance with a
preferred embodiment of the present invention;
[0022] FIG. 3 is a flowchart depicting a process for loading
database information containing person-by-person information from
one storage location to another in accordance with a preferred
embodiment of the present invention;
[0023] FIG. 4 is a block diagram of a data conversion method
according to a preferred embodiment of the present invention;
[0024] FIG. 5 is a filter mask according to a preferred embodiment
of the present invention;
[0025] FIG. 6 is a detailed graphical representation of a preferred
embodiment of the data structure of the .tvd files in the
database;
[0026] FIG. 7 is a graphical representation of the viewing catalog
for a three-week period;
[0027] FIG. 8 is a flowchart of a method 800 for using a graphical
user interface to analyze the records in the database using a
preferred embodiment of the present invention;
[0028] FIG. 9 is a screen shot of cross tabulation data extracted
from a media-related record in a database file according to a
preferred embodiment of the present invention;
[0029] FIG. 10 is a screen shot of an icon for accessing data
contained in a database file according to a preferred embodiment of
the present invention;
[0030] FIG. 11 is screen shot of a line graph representing data
contained in a database file according to a preferred embodiment of
the present invention;
[0031] FIG. 12 is a detailed graphical representation of a
media-related database structure according to an alternative
preferred embodiment of the present invention;
[0032] FIG. 13 is a flow chart depicting an optimization method
according to a preferred embodiment of the present invention;
[0033] FIG. 14 is a block diagram of a scoring calculation method
according to a preferred embodiment of the present invention;
[0034] FIG. 15 is a tabular representation of a viewing history for
a series of advertising spots;
[0035] FIG. 16 is tabular representation of frequency values for a
given number of exposures to an advertising spot;
[0036] FIG. 17 is a tabular representation of age range index
values for a group of audience members;
[0037] FIG. 18 is a tabular representation of income range index
values for a group of audience members;
[0038] FIG. 19 is a tabular summary of various techniques for
valuing exposures to an advertising message;
[0039] FIG. 20 is a tabular comparison of comparative average
exposure frequencies according to two different plans or
schedule;
[0040] FIG. 21 is a graphical representation of two different
exposure plans or schedule for an advertisement;
[0041] FIG. 22 is a tabular representation of effective frequency
for a simple advertising campaign;
[0042] FIG. 23 is a tabular representation of sample index
valuation for a given number of exposures to an advertisement;
[0043] FIG. 24 is a graphical comparative representation of a
number of exposure valuation models;
[0044] FIG. 25 is a graphical comparative representation of total
exposure valuation using several different models;
[0045] FIG. 26 is a tabular representation of exposure valuation
for various frequencies as applied in several different models;
[0046] FIG. 27 is a tabular representation of scoring exposure
values as applied in several different models;
[0047] FIG. 28 is a graphical representation of change in frequency
based on choosing alternatives to a base plan or schedule;
[0048] FIG. 29 is a tabular representation of frequency exposure
valuation for a series of multiple exposures to a given
advertisement;
[0049] FIG. 30 is a graphical representation of the decay of
influence resulting from the passage of time;
[0050] FIG. 31 is a graphical representation of an index for
logarithmic influence;
[0051] FIG. 32 is a graphical representation of influence to
advertising based on exposure frequency;
[0052] FIG. 33 is a graphical representation of the average
influence index values for three groups of audience members;
[0053] FIG. 34 is a graphical representation an influence index
using two different exposure frequencies;
[0054] FIG. 35 is a tabular representation of a scoring example
using a time weighted effective frequency model;
[0055] FIG. 36 is a graphical representation of an influence index
for a given audience member;
[0056] FIG. 37 is a tabular representation of index values based on
audience age and gender information;
[0057] FIG. 38 is a tabular representation of index values based on
income information;
[0058] FIG. 39 is a tabular representation of index values based on
the size of the county where the audience member resides;
[0059] FIG. 40 is a tabular representation of index values for
exposure recency;
[0060] FIG. 41 is a tabular representation of scoring using index
values and models according to a preferred embodiment of the
present invention;
[0061] FIG. 42 is a tabular representation of indices for three
advertising alternatives using a scoring model according to a
preferred embodiment of the present invention;
[0062] FIG. 43 is a tabular representation of values for a series
of advertising spots for optimization purposes; and
[0063] FIG. 44 is a graphical representation of the optimization
information presented in FIG. 43.
BEST MODE FOR CARRYING OUT THE INVENTION
[0064] The various preferred embodiments of the present invention
below are described in connection with the person-by-person data
gathered and distributed by Nielsen Media Research Service. This
approach has been selected to more clearly and distinctly explain
the various embodiments of the present invention. However, although
the present invention is described in the context of television
viewing, it is important to note that person-by-person data for any
type of media may be utilized with various preferred embodiments of
the present invention. For example, various embodiments of the
present invention are contemplated to address readership
information for magazines or newspapers as well as browsing
information for individuals accessing web pages on the World Wide
Web. The advertising optimization techniques described herein are
equally effective for media advertising purposes in all other media
as well.
[0065] In addition, although the present invention is described in
the context of the Nielsen data as used in the United States, many
other countries also have data collection services, which supply
minute-by-minute, person-by-person data, rather than the mid-15
minute information supplied by Nielsen. The system and methods of
the present invention are equally effective for these types of
databases.
[0066] For those individuals who are not well versed in the
generation and use of person-by-person data such as the data
associated with the Nielsen Media Research Service, the Overview
section below presents many of the concepts that will help to
understand the invention. Those who are generally familiar with the
present state of television ratings systems and analysis may
proceed directly to the detailed description section below.
[0067] 1. Overview
[0068] Television viewing for the population of the United States
is estimated by A. C. Nielsen Company (Nielsen) based on viewing
logs generated from a sample of 5,000 households, with a total of
about 15,000 sample members living in those households. Using
specialized equipment attached to televisions in the homes, and
communicating with these devices using telephone line connections;
Nielsen accumulates data. The Nielsen data describes the viewing
choices of each of the household members in a time segment format.
This viewing information is packaged and sold as a service by
Nielsen to television stations, network programmers, advertising
agencies, marketing research groups, universities, and other
interested individuals.
[0069] On a weekly basis, Nielsen supplies to subscribers of their
services person-by-person data files that detail the television
viewing choices for each of the sample members in the sample
population. The data contained in these files indicate whether or
not each of the sample members was watching television during the
midpoint of each 15-minute period during the week. If the sample
member was watching television, the data also indicate which
television program they were watching. Also delivered with the data
is additional information describing the demographic
characteristics of the households and sample members living in each
household, such as geographic territory, household income,
education, age, and gender of the members living in the household.
The typical size of a data file for one week is about 35
Mbytes.
[0070] On a regular basis, Nielsen modifies the overall sample
population profile by adding new households or dropping certain
households from the sample. A household can be included in the
sample for as long as two years, after which time the household
will be removed from the sample population. However, for various
reasons such as relocation, death, and unreliability, households
will often remain in the sample for only a few weeks or months.
Also, any given sample household and/or sample member(s) in the
sample population can be judged by Nielsen to be out of tab for a
day. This means that the television viewing data for a particular
household or member for that day are not reliable for estimation
and reporting purposes. Similarly, those households and members
that have apparently reliable data for a given day are termed
in-tab by Nielsen.
[0071] Nielsen assigns each of the sample households a globally
unique household number that is used to track the data for that
household as long as the household remains in the sample.
Similarly, each household member is assigned a unique person number
within each household. Thus, even though the data files are
delivered to subscribers in separate files each week, the
television viewing habits of sample members can be tracked for each
and every week in which they are included in the sample.
[0072] Although people visiting the sample households and watching
television in a sample household are logged and identified in the
data, they are not assigned unique person numbers. Therefore, the
viewing history for household visitors cannot be accurately tracked
from one day or week to the next. Nielsen does, however, provide
information regarding the age and gender of each visitor in a
Nielsen household.
[0073] As previously mentioned, although this very significant body
of data are available for use by businesses and organizations that
hope to benefit by analyzing it, the limitations of the present
analysis tools and techniques have been too cumbersome to overcome.
The two main factors which in the past have made the Nielsen
person-by-person data difficult to work with are the tradeoff
between database size and system performance, and the problems
associated with the unification of the sample data. In addition,
several other difficulties with accessing and analyzing the Nielsen
data have been noted. These various problems are explained briefly
below.
[0074] Many types of data analysis require the use of many weeks,
months, or, in some cases, years of data. Patterns of program and
network loyalty, for example, cannot be effectively analyzed
without a large quantity of data which spans periods that are at
least as long as the time required to established these patterns.
In the case of the Nielsen data, each additional week of
person-by-person data increases the size of the Nielsen database by
approximately 35 Mbytes.
[0075] As the size of the person-by-person database increases,
system-level processing constraints and limitations become a
significant concern. If, for example, a person wanted to select
only those members of the sample population who fell into a
particular demographic group, then the entire database for all
weeks would need to be filtered using the specific demographic
criteria selected. Using the Nielsen data, this task can be
daunting.
[0076] In the context of analyzing the Nielsen data, the principle
of sample unification refers to the process of correlating the
individual data elements from multiple sets of data. The viewing
data from each viewing day and week for each sample member, for
example, needs to be matched or correlated for analysis. If a
particular query requires data for 10 days of viewing spanning a
four-week calendar period, then only those sample members who were
both in the sample and in-tab for each of the 10 days should be
included in the sample. For limited data sets and moderate sample
sizes, the sample unification process is manageable using
conventional database techniques. For data sets of the size
available from Nielsen, the unification task becomes daunting one
without the creation of specialized tools as described herein.
[0077] The most ambitious studies using Nielsen data previously
undertaken by independent researchers and practitioners have
spanned only a very few weeks, generally two to four weeks. These
researchers have invariably noted the difficulties that they have
encountered with the sheer size and cumbersome nature of the data.
Periodically, Nielsen processes requests for specialized studies
made by subscribers of their services that include more data than
this. However, even these special studies still generally span a
limited number of weeks of data, require the use of the substantial
computing resources available to Nielsen, and may require several
weeks of calendar time to generate the results.
[0078] Other problems related to the size of the Nielsen data files
are particularly troublesome for organizations with multiple
geographic locations that need to access the Nielsen data files.
Organizations with offices spanning large geographic areas require
duplication of the database in each location to avoid frequent
remote retrieval of large quantities of data. Weekly updates to
each database with 35 Mbytes of data would become prohibitively
difficult. In addition, if such a data service became widely
popular, the task of delivering and installing data for numerous
clients using conventional database management systems would also
become unreasonably difficult for a data vendor using existing
technology.
[0079] In addition, users frequently need access to the data using
desktop and laptop computer systems. This requires duplicating
portions of the database on other computer systems. For example, if
a user needs stand-alone access to three months worth of Nielsen
data, they would have to install over 250 Mbytes of data into a new
environment.
[0080] Yet another problem becomes apparent when using the Nielsen
data for demographic studies and analysis. One mode of inquiry into
the Nielsen data could be termed "browsing." In this mode a user
interactively interrogates the data, often with a given inquiry
being based on a previous inquiry. A user would, for example, look
at some type of viewing analysis for a selected demographic group,
then alter the selected demographic and examine the same analysis
for comparison. The data analysis system, then, needs to be able to
select a subset of the sample based on the demographic parameters
chosen by the user. This, of course, is not a difficult task to
perform for many database management systems. The SQL database
query language, for example, supports these types of queries. But
to do so with sufficient speed to allow interactive browsing is not
possible at the current time with widely available computer systems
and traditional DBMS software. The data sets are simply too large
and the queries too complex.
[0081] Finally, the physical structure of the records in the
Nielsen data files can also introduce unnecessary complexity into
the analysis process. Nielsen supplies data to their subscribers in
a record-oriented format. Six different record types are used to
detail the person-by-person data and each of these six record types
is explained briefly below.
[0082] Calendar record (record type 0). The calendar record
identifies the broadcast week for all other records. One calendar
record is provided for each week's data.
[0083] Example:
[0084]
005297051297051397051497051597051697051797051809705120970513097
05140970515097051609705170970518
[0085] This calendar record indicates that the week of data
included in the data set begins on May 12, 1997 and ends on May 18,
1997.
[0086] Classification data record (record type 1). The
classification data record describes each household in the sample
in terms of income, education of the head of the household, time
zone, etc. This record also specifies the age and gender of each
household member and visitors to the household during the week. The
Nielsen data file typically includes 5000 records per week
[0087] Example:
[0088] 100040520200034117248342520222222213109605305
A01Q02Y03J04A00
[0089] The classification data record shown above describes
household number 200034. It indicates that the household was added
to the Nielsen sample on the 53.sup.rd day of 1996, that the
household was in-tab for all seven days of the week, and that the
household for the week includes four household members and one
visitor. Each household data item in each classification record
will be translated to binary form. The income indicator "5" in the
record above, for example, is translated to the binary number
"00100000." Similarly, the age and gender data for each person in
the record is also translated to binary form and assigned to the
corresponding attribute in an object created for the person.
[0090] Program data lead record (record type 2). The program data
lead record describes each quarter hour of programming broadcast
during the week including program name, episode name, air date and
time, program genre, etc. Typically, 1500 program data records are
used to describe all programming broadcast in a given week
[0091] Example:
[0092] 20453 1890800003020SEINF
052501SEINFELD2100NCS09480128599
[0093] The sample program data lead record shown above indicates,
among other things, that the Seinfeld show airs on NBC at 2100
hours (9:00 PM), it's a situation comedy, and that the program run
length is 30 minutes. Dates, times, and quarter hour values in the
program data item are converted from an ASCII representation as
shown above to a binary representation, and assigned to
corresponding attributes in a program object.
[0094] Program data continuation record (record type 3). Specifies
those households and household members who were logged as viewing
each quarter hour of network programming broadcast during the week
as described by the program data lead record. There are 50,000
program data continuation records supplied in the Nielsen data per
week.
[0095] Example:
[0096] 3045302000340000200034Y0302000740000200074X01
02000741020200251
000200251T0102003410000200341M0102003820000200382X010200382P02
[0097] This record indicates some of the households and people who
were watching Seinfeld at 9:00 PM on NBC. As indicated by the 10
character string "200034000" (the "viewing event string") in the
record, one of the households was number 200034. In that household,
as indicated by a 10 character viewing event string in the record
"200034Y03", person number Y03 was also watching the program.
[0098] Usage data lead record (record type 4). The usage data lead
record identifies by sequence number each quarter hour during the
broadcast week. There are 672 records per week contained in the
Nielsen data files.
[0099] Example:
[0100] 400660205222200274804115
[0101] This usage data lead record assigns the sequence number 0066
to the quarter hour which begins at 10:00 PM on the 2.sup.nd day of
the week (Monday). 2,748 households in the sample were using their
televisions at that time, and 4,115 people were watching television
in those households. The date/time in each usage data lead record
is noted and used in reading each subsequent usage data
continuation record.
[0102] Usage data continuation record (record type 5). Specifies
those households and household members who were logged as using the
television during each quarter hour as noted by the usage data lead
records. There are 200,000 usage data continuation records per week
supplied in the Nielsen data files.
[0103] Example:
[0104]
5006602000340000200034Y030200034J0402000830000200083W010200105
000200105W010200105I0202002010000200201I0402002510000200251IU02
[0105] This usage data continuation record identifies some of the
2,748 households and 4,115 people in the sample who were watching
television on the 2.sup.nd day of the week at 10:00 PM. One of the
households was number 200034, as indicated by the 10 character
viewing event string "200034000" in the record. In that household,
person number Y03 was watching, as indicated by a 10 character
"200034Y03" viewing event string in the record. The specific
program being watched is not specified. It may have been a network
program. Alternatively, it may also have been a cable channel or a
broadcasting station unaffiliated with one of the networks.
[0106] Two separate records are used to describe television viewing
in the Nielsen data files, a program data record and a usage
record. The program data record indicates all those persons and
households in the sample who viewed a particular network program.
Viewing of non-network programming is not indicated in the "program
data" record. The "usage" record indicates television usage by
person and household. With this data, to find all instances where
sample members watched non-network programming, those instances in
which the program data indicates the household members were
watching network programming must be subtracted from the usage
entries. The usage entries remaining after this subtraction are the
non-network viewing entries.
[0107] 2. Definitions
[0108] Homes Base. Home base is defined as the total number of
homes in the United States with one or more television sets.
Approximately 95 million homes fall into this category. This
figure, on a percentage basis, includes more than 98% of all
homes.
[0109] Prime Time. The federal government defines prime time as
those evening hours during which the television networks are
allowed to broadcast their programming. These hours are from 7:00
PM to 11:00 PM on Sunday, and from 8:00 PM to 11:00 PM
otherwise.
[0110] Households Using Television (HUTS). HUTS is the total number
of homes with televisions turned on at a given time. During prime
time this number is often over 60 million. The percentage of all
homes with television sets which had those sets turned on is
referred to as "percent HUTS" or "HUT rating." If the homes base
was 95 million and HUTS was 60 million, then the HUT rating would
be
HUT rating= homes watching television/homes with television=60
million/95 million=63.2% HUTS
[0111] This value is consistently referred to as simply HUTS rather
than percent HUTS or HUT rating. True HUTS (in millions of homes)
is rarely used. This convention is adopted herein as well. In every
case where the term "HUTS" is used, this will actually refer to HUT
rating. HUT levels typically peak at over 60% during prime time and
can be less than 20% between 7:00 and 10:00 a.m. during the summer.
HUTS typically bottoms out at about 4:00 a.m. on weekdays at about
7%.
[0112] Household Rating. Household rating is defined as the portion
of all homes having televisions sets which had those sets tuned to
a particular show. Assuming 95 million televisions in the nation
and 15 million are tuned to a particular show then
rating= homes watching a show/homes base=15 million/95
million=15.8%
[0113] Demographic rating is similar to household rating, but this
figure is calculated using the number of people in a particular
demographic group who saw the show divided by the number of people
in the population for that demographic group. This figure is used
as another indicator used for decision-making based on
demographics.
[0114] Share. Share is defined as the portion of homes with
television sets on which were tuned to a particular show. If 60
million homes had televisions turned on, as in the example above,
and 15 million were watching a particular show, then
share= homes watching a show/homes watching television=15
million/60 million=25%
[0115] Rating and share are related through HUTS.
rating= HUTS.times.share=63.2%.times.25%=15.8%
[0116] Homes Delivered. Advertising effectiveness is sometimes
based on homes delivered to an advertiser. This value is defined as
the rating for a show multiplied by the total number of homes with
television sets. The homes delivered for a show with a 15.8% rating
would be
homes delivered= homes base.times.rating=95
million.times.15.8%=15.0 million
[0117] This is equivalent to the total number of homes with their
television sets turned on multiplied by the share. Assuming that 60
million homes were watching the show
homes delivered=homes base.times.HUTS.times.share=95
million.times.63.2%.times.25%=15.0 million
[0118] Special rules defined by Nielsen apply for computing these
types of household-based measurements for households with multiple
television sets during those times when the television sets in the
household are tuned to different channels. See Nielsen Media
Research [1994] for a detailed description of these rules.
[0119] Viewers Per Viewing Household (VPH). VPH is defined as the
number of viewers of television averaged over all households
watching television and varies by half-hour and by show. Some shows
tend to have larger groups of people watching than other shows.
This number by definition is never less than one, and is rarely
over two.
[0120] Impressions. Advertisers often speak in terms of
impressions. An impression is defined as one person viewing either
one show or one advertisement one time. It can be calculated using
homes delivered and VPH. Assuming that for a particular show, the
average number of people watching per home is two, then impressions
is the product of homes delivered and VPH.
impressions= homes delivered.times.VPH=homes
base.times.rating.times.VPH=9- 5 million.times.15.8%.times.2=30.0
million
[0121] 3. Detailed Description
[0122] As explained above, a computer-based system according to a
preferred embodiment of the present invention includes four main
components: a database mining engine (DME); a DME database; an
optimization mechanism; and a user interface which controls the
system and allows a user to manipulate and analyze the data in the
DME database by using the DME. Taken together, these components
offer a powerful tool for manipulating and analyzing Nielsen viewer
data for decision-making purposes. Further, since the various
embodiments of the present invention are designed to be used in a
computer-based environment, a suitable computer system is
necessarily a part of the present invention. Each of these main
components will now be described in greater detail.
COMPUTER SYSTEM
[0123] Referring now to FIG. 1, a computer-based system 100 for
advertising optimization in accordance with the most preferred
embodiment of the present invention includes an IBM PC compatible
computer. However, those skilled in the art will appreciate that
the methods and apparatus of the present invention apply equally to
any computer system, regardless of whether the computer system is a
complicated multi-user computing apparatus or a single user device
such as a personal computer or workstation. Computer system 100
suitably comprises a processor 110, main memory 120, a memory
controller 130, an auxiliary storage interface 140, and a terminal
interface 150, all of which are interconnected via a system bus
160. Note that various modifications, additions, or deletions may
be made to computer system 100 illustrated in FIG. 1 within the
scope of the present invention such as the addition of cache memory
or other peripheral devices. FIG. 1 is presented to simply
illustrate some of the salient features of computer system 100.
Those skilled in the art will recognize that there are many
possible computer systems which will be suitable for use with the
present invention.
[0124] Processor 110 performs computation and control functions of
computer system 100, and comprises a suitable central processing
unit (CPU). Processor 110 may comprise a single integrated circuit,
such as a microprocessor, or may comprise any suitable number of
integrated circuit devices and/or circuit boards working in
cooperation to accomplish the functions of a processor. Processor
110 suitably executes an object-oriented computer program 122
within main memory 120.
[0125] Auxiliary storage interface 140 allows computer system 100
to store and retrieve information from auxiliary storage devices,
such as magnetic disk (e.g., hard disks or floppy diskettes) or
optical storage devices (e.g., CD-ROM). One suitable storage device
is a direct access storage device (DASD) 170. As shown in FIG. 1,
DASD 170 may be a floppy disk drive which may read programs and
data from a floppy disk 180. It is important to note that while the
present invention has been (and will continue to be) described in
the context of a fully functional computer system, those skilled in
the art will appreciate that the mechanisms of the present
invention are capable of being distributed as a program product in
a variety of forms, and that the present invention applies equally
regardless of the particular type of signal bearing media to
actually carry out the distribution. Examples of signal bearing
media include: recordable type media such as floppy disks (e.g.,
disk 180) and CD ROMS, and transmission type media such as digital
and analog communication links, including wireless communication
links.
[0126] Memory controller 130, through use of a processor (not
shown) separate from processor 110, is responsible for moving
requested information from main memory 120 and/or through auxiliary
storage interface 140 to processor 110. While memory controller 130
is shown as a separate entity for purposes of explanation, those
skilled in the art understand that, in practice, portions of the
function provided by memory controller 130 may actually reside in
the circuitry associated with processor 110, main memory 120,
and/or auxiliary storage interface 140.
[0127] Terminal interface 150 allows system administrators and
computer programmers to communicate with computer system 100,
normally through programmable workstations. Although the system 100
depicted in FIG. 1 contains only a single main processor 110 and a
single system bus 160, it should be understood that the present
invention applies equally to computer systems having multiple
processors and multiple system buses. Similarly, although the
system bus 160 of the preferred embodiment is a typical hardwired,
multidrop bus, any connection means that supports bi-directional
communication in a computer-related environment could be used.
[0128] Main memory 120 suitably contains an operating system 122, a
graphical user interface 125, a Database Mining Engine (DME)
database 126, a Database Mining Engine (DME) 127, a data conversion
mechanism 128, and an advertising optimization mechanism 129.
Operating system 122 in memory 120 is used to control the
functional operation of system 100. Graphical user interface 125 in
memory 120 provides access for a user of system 100, allowing the
user to access the various features of system 100. DME database 126
is a customized version of a previously created database that is
optimized for access by DME 127 via graphical user interface 125.
DME 127 is a specialized database management system (DBMS) which is
optimized to search, manipulate, and analyze person-by-person
records in a database format. DME 127 uses a customized set of
filters to access the data contained in DME database 126 to
formulate responses to queries from a user of system 100.
[0129] Advertising optimization mechanism 129 allows a user of
system 100 to interactively create, score, analyze, and compare
various advertising-related decisions, using the data contained in
DME database 126. Although shown as separate components for this
example, the various components shown in memory 120 may be provided
separately or, alternatively, may be individual parts of a single
software program. The various components loaded into memory 120 are
typically loaded into memory 120 from a secondary storage location
such as DASD 170. The term "memory" as used herein refers to any
storage location in the virtual memory space of system 100.
[0130] It should be understood that main memory 120 does not
necessarily contain all parts of all mechanisms shown. For example,
portions of operating system 122 may be loaded into an instruction
cache (not shown) for processor 110 to execute, while other files
may well be stored on magnetic or optical disk storage devices (not
shown). In addition, although Database Mining Engine (DME) 127 is
shown to reside in the same memory location as DME database 126 and
operating system 122, it is to be understood that main memory 120
may consist of multiple disparate memory locations (e.g. backside
cache, look-aside cache, etc.).
DME AND DME DATABASE
[0131] 1. Introduction
[0132] The following section describes some of the most salient
features of DME database 126, DME 127, and the associated
techniques and tools used in preparing the Nielsen
television-related viewing data (Nielsen data) for use in system
100. In order to more quickly and efficiently process the large
volume of data contained in the Nielsen data files, the generally
accepted concepts of database design and manipulation so prevalent
today must be discarded or modified. This is simply because the
various relational and hierarchical database models in use today
are too unwieldy for manipulating large data files with any
significant speed, absent very specialized and expensive computer
hardware. Although not preferred, the methods of the present
invention can be practiced with other, less efficient models.
[0133] The unique format of DME database 126, combined with the
functional aspects of DME 127, overcomes several limitations of
conventional database and data processing techniques which tend to
reduce the performance of most data analysis systems to
unacceptable levels. Although the various preferred embodiments of
the present invention are presented and described in the context of
television viewing, other types of data may be manipulated and
analyzed in a similar fashion. It should be noted that the concepts
and techniques of the present invention are equally applicable to
tracking and analyzing the behavior of a sample population for
visitors to web pages on the World Wide Web. Similarly, information
about the readership populations for magazines and newspapers could
also be manipulated and analyzed by applying various preferred
embodiments of the present invention. Indeed, any advertising
firm/agency, business, or other organization that wishes to track
large quantities of information regarding various sample
populations can successfully implement the various techniques and
methods described herein.
[0134] A system 100 according to various preferred embodiments of
the present invention has the following significant advantages: the
ability to add, on a weekly basis, large quantities of data to the
existing user databases; a way to easily move relevant portions of
existing databases from location to location (such as from a
central server to a laptop computer); the ability to retrieve large
blocks of data from the database, organize the data in memory, and
analyze the data; the ability to filter the data according to user
selected demographic criteria; and retrieve information for the
same sample members across multiple weeks.
[0135] The various capabilities listed above are a direct product
of the unique design of DME database 126 and the techniques
associated with manipulating the data contained in DME database
126. The design of DME database 126 is performance driven and, for
at least one preferred embodiment, is specifically designed to
efficiently access the Nielsen data. Using most standard computers,
the performance of a general purpose DBMS will typically be
inadequate for interactive analysis when manipulating the hundreds
of Mbytes of data that comprise the Nielsen data. Recognizing this,
a custom DBMS (i.e. DME database 126 and DME 127) can be created to
take advantage of the specific characteristics of the data (in this
case, television viewing data prepared by Nielsen). The
organization and manipulation methods and techniques for accessing
DME database 126 are described below.
[0136] 2. Detailed Description of DME Database
[0137] DME Database Organization
[0138] A DME database 126 according to a preferred embodiment of
the present invention is capable of spanning many weeks and is
composed of .tvd discrete files, one file corresponding to each
week of the Nielsen data. The name assigned to each of the files in
DME database 126 is the date of the Nielsen data contained within
that file. For example, data for the week ending Jul. 28, 1997 is
contained in a file with the name 19970728.tvd. TVD is an acronym
for Television Viewing Data and is used to identify all files with
a format suitable for use with the present invention. To add an
additional week of Nielsen data to DME database 126, the .tvd file
that contains that week's viewing data is simply placed into the
directory for DME database 126 along with all other .tvd files.
This feature of the present invention makes it very easy to keep
DME database 126 up to date.
[0139] A user or system administrator can create a copy of all or a
selected portion of DME database 126 by copying some or all of the
.tvd files to another memory storage location, and then, by using a
command that can be accessed through graphical user interface 125
described below, direct system 100 to access the new database
location. No other database installation process is required.
[0140] DME Database File Format
[0141] The .tvd data files contained in DME database 126 can be
considered object-oriented for several reasons. First, the various
components of the Nielsen data are treated as a group of various
objects, i.e. a household object, a person object, a television
program object, etc. Accordingly, all of the data for each discrete
object, such as a household, person, or television program object,
is located contiguously in the .tvd file.
[0142] For example, the data describing a particular person's age,
gender, and person number are physically adjacent in the individual
database file, rather than as columns in a relational table. In
addition, the length and relative byte position of each data
element for each object in the database file is the same as the
required length and byte position of those same data elements in
memory 120. Further, the relative positions of data in the file and
in memory 120 are the same. A region of memory 120 is allocated for
loading a sets of person objects (or program objects, or household
objects). Memory 120 is sized according to the type of object being
loaded and number of objects in the collection.
[0143] Given this memory allocation, the data is loaded in a binary
fashion from the Nielsen data file into memory 120. During this
loading process, data attributes are ignored. The first byte of
person object data for the person collection in the file, for
example, is loaded into the first byte of allocated memory. The
second byte of data is then loaded into the second byte of
allocated memory, etc. This process is both fast and reliable. It
is important to note that the data are not loaded as objects, but
once the data is loaded into memory 120, system 100 can operate on
the data as objects. Third, data items are retrieved from the
database as objects and collections of objects, rather than as
discrete data elements which are assembled into objects in memory.
Finally, the data for similar objects, such as people in the same
household, or programs of the same day, are also located
contiguously in the file.
[0144] This unique database structure allows for binary data
transfer of large blocks of objects from disk to memory 120. The
objected-oriented database management software (DME 127), requires
memory based data objects for processing and can begin operating on
the .tvd data immediately after retrieval from DME database file
126.
[0145] DME Database File Creation
[0146] Although the data received from Nielsen is very valuable for
analysis and advertising purposes, the format of the data does not
readily lend itself to quick and efficient manipulation. For this
reason, the various preferred embodiments of the present invention
will read the data from the magnetic tapes supplied by Nielsen,
reformat it, and store it in DME database 126 using .tvd files. The
data conversion process is detailed in FIG. 2.
[0147] Referring now to FIG. 2, a process 200 in accordance with a
preferred embodiment of the present invention for converting data
from a first data format (i.e., the Nielsen format) to a second
format (i.e., the .tvd format) is illustrated. As shown in FIG. 2,
process 200 generally involves organizing the input data from the
standard format (in this case, the Nielsen format) into the format
required for object-oriented processing, then writing this memory
data in binary form to individual .tvd files within DME database
file 126. The basic steps for this process are: allocate blocks of
memory 120 (step 210); assign these memory blocks to arrays of
objects, such as arrays of person objects or program objects (step
220); read the data supplied by Nielsen and assign values to the
object data elements in memory, such as age, or program name (step
230); and, write the blocks in binary form from memory 120 into a
newly created DME database .tvd file (step 240). It is important to
note that there is no requirement to locate all blocks of memory
120 in a contiguous fashion. Blocks of memory 120 may be allocated
as needed, where needed to accommodate the input data.
[0148] Referring now to FIG. 3, a process 300 in accordance with a
preferred embodiment of the present invention for accessing DME 127
is illustrated. When the data in DME database 126 is being used for
analysis, the direction of data transfer is reversed. First, blocks
of memory 120 are allocated and the blocks of memory 120 are
assigned to arrays of objects (step 310); the blocks of data are
read from the .tvd data files in DME database 126 in binary form
into the allocated memory blocks (step 320); and then DME 127 can
access the television viewing data. This type of data retrieval is
not possible using conventional database systems because the binary
representation of the data in a typical database is typically not
the same as the data in memory 120. As explained above, for the
various preferred embodiments of the present invention, the data
storage format is identical.
[0149] Note, however, that if the structure of the various data
objects in DME database 126 is modified to accommodate expanded
types of analysis, then the .tvd database files also need to be
modified to reflect those changes. As noted above, data is binary
loaded from database file 126 into memory 120. Portions of memory
120 are sized according to the type of object being loaded and the
number of objects in the collection. Using a database that is
composed of 52 discrete weekly .tvd files, the assumed size of
person objects in each of the 52 files is identical. Person objects
in collections loaded from the week 1 file are the same size as
person objects from the week 10 file.
[0150] For example, if the size of person objects in the Nielsen
data is expanded by adding additional attributes, such as an
occupation attribute or a head-of-household flag, this change would
also need to be reflected in the size of the person objects in each
of the 52 weeks of data. All 52 files would need to be recreated
with the newly resized person objects. The design of DME database
126 does not provide the data type independence of many
commercially available database management systems in which the
representation of the data in the database is independent of the
representation of the data in memory. However, by mirroring the
data in both locations, a significant speed advantage is
recognized.
[0151] As previously mentioned in the Overview section, the records
in the Nielsen data files are converted for use with the preferred
embodiments of the present invention. Each of the six supplied
files plays a part in creating DME database 126 and the following
conversion details are performed by DME 127. Once again,
information in the calendar record is read but is not entered into
the TVD database file. The data in the calendar record is used only
to validate dates in other record types.
[0152] Each household data item in each classification record is
translated to binary form as described in the section "Sample
Filtering". This binary form of the data item is then assigned to
the corresponding attribute in the household object created for the
household. The income indicator "5" in the record above, for
example, is translated to the binary number "00100000." Similarly,
the age and gender data for each person in the classification
record is also translated to binary form and assigned to the
corresponding attribute in the person object created for the
person. As Household and Person objects are created, they are added
to their respective Group collections.
[0153] Each program data item in each program data lead record is
assigned to the corresponding attribute in the program object
created for the program. Dates, times, and quarter hour values in
the program data item are converted from an ASCII representation as
shown above in the Overview section to a binary representation, and
assigned to the corresponding date/time attributes in the program
object. As Program objects are created, they are added to the
Program Group collection.
[0154] Prior to the step in the creation of DME database 126 where
the program data continuation records are read, a Viewing Index is
created, as described in the section "Viewing Data", and
illustrated in the figure showing "Database Structure."
[0155] After creating the Viewing Index, for each network viewing
event in each program data continuation record, the Household or
Person object referred to in the event is found in the Household
Group or Person Group collections. As described in the section
"Viewing Data," the memory location in the viewing data memory for
this household/person and date/time is identified, and a notation
is made indicating that the household/person viewing the network
program at the indicated date/time. The date/time in each usage
data lead record is noted and used in reading each subsequent usage
data continuation record.
[0156] Using the date/time from the usage lead record, for each
viewing event in each usage data continuation record, the Household
object or Person object referred to in the event is found in the
Household Group or Person Group collections. As described in the
section "Viewing Data", the memory location in the viewing data
memory for this household/person and date/time is identified. At
this point a notation may be made to the record, conditioned on the
presence or absence of a preexisting notation:
[0157] 1) If a notation has already been made in this location
based on a viewing event string in a Program Data Continuation
record indicating that the household or person was watching network
television, then no further notation is made. It can be assumed
that the usage viewing event refers to this network viewing.
[0158] 2) If no notation is found, then it can be assumed that the
household or person was viewing non-network television, and a
notation is made accordingly.
[0159] After processing all Nielsen data records for a given week,
the resulting memory objects are written to disk as a TVD database
file. Each memory location is written in binary form in sequence:
first the viewing index is written. This index includes the offset
value described in the section "Viewing Data". Following the index,
all household objects are written, followed by all person objects,
and program objects. Finally, the actual viewing data is written,
all in binary form.
[0160] As described above, in the Nielsen data files, two separate
records are used to describe television viewing. One is a program
data record, and the other is a usage record. The program data
record indicates all those persons and households in the sample who
viewed a particular network program. Viewing of non-network
programming is not indicated in the "program data" record. The
"usage" record indicates television usage by person and household.
With this record arrangement, to find all instances where sample
members watched non-network programming, those instances in which
the program data indicates the household members were watching
network programming must be subtracted from the usage entries. The
usage entries remaining after this subtraction are the non-network
viewing entries.
[0161] It is important to note that this technique of assigning
non-network viewing to those individuals who are viewing television
but not viewing network programming will result in occasional
over-counting of non-network program viewing. This occurs when the
television stations in certain markets delay the broadcasting of
regularly scheduled programming. Av viewer can watch the program at
the delayed time. Nielsen will attribute the viewing for the
program at the regularly scheduled time, but also note television
usage at the delayed time. When this occurs, the network television
viewing audience counts will be accurate, but the non-network
audience counts will be inflated.
[0162] Referring now to FIG. 12, an alternative preferred
embodiment for storing information in database 126 is illustrated.
Rather than allocate a fixed set of memory locations for each
person/viewing time as shown in FIG. 6, memory space can be
conserved by allocating memory only for changes in viewing status.
Each viewing record is a variable length record which tracks each
change in viewing status over a given period of time. For example,
referring to record number 0 in FIG. 6, the viewing status for the
individual represented by record 0 changed between 6:23 and 6:38.
This change is illustrated in FIG. 12 by a single entry of
"6:23<blank>" which indicates that from the beginning of the
period covered by the viewing data through 6:23 p.m., the
individual represented by this record was not watching television.
The next entry is "7:23, n" meaning that the viewer was watching
the "n" network beginning from the previous entry through 7:23 p.m.
The last entry for each for the records is one for 11:53 p.m., the
last time entered for the period. Using these entries as an
end-of-file (EOF) delimiter, the end of each variable length record
can be detected.
[0163] Data Conversion Mechanism
[0164] Referring now to FIG. 4, a sample data conversion mechanism
128 and process adapted for use with the present invention are
illustrated. A data conversion mechanism 128 according to a
preferred embodiment of the present invention is a
computer-implemented process for converting person-by-person
media-related data from a first data format to a second data
format. The data conversion mechanism will use Processor 110 to
execute the process and memory 120 as a storage location in order
to convert the data. The process of creating a .tvd data file for
DME database 126 involves reorganizing a weekly set of data
delivered by Nielsen into the form required for object-oriented
processing, then writing this memory data in binary form to DME
database 126.
[0165] Referring now to FIG. 4, the steps for performing data
conversion via process 400 are described.
[0166] 1) Allocate blocks of memory 120 which are sufficiently
large to accommodate the data in the week being processed, and
assign these memory blocks to arrays of household objects 454,
person objects 456, program objects 458, and viewing data 460.
Because it is unknown at this stage of the process exactly how many
programs, households, etc. are include in the Nielsen data for the
week, it is likely that portions of each of these memory blocks 120
will not be used, and will not be written to the completed .tvd
database file 126.
[0167] 2) Read all classification records.
[0168] a) Read a classification record from Nielsen data 430.
[0169] b) Select the next household object in the household object
array 454 for this data.
[0170] c) Translate to binary form the household attributes in the
classification record. Each household data item in each
classification record is translated to binary form as described in
the section "Sample Data Filtering" below. This binary form of the
data item is then assigned to the corresponding attribute in the
household object selected for that given household. The income
indicator "5" in the record above, for example, is translated to
the binary number "00100000".
[0171] d) Increment the household object array counter in Household
Group 440 to indicate that the household has been added to
household object array 454.
[0172] e) For each person in the classification record, select the
next person object in person object array 456.
[0173] f) Translate to binary form the person attributes of age and
gender data, and assign to the corresponding attribute in the
person object selected for the person.
[0174] g) Increment the person object array counter in Person Group
442 to indicate that the person has been added to person object
array 456.
[0175] 3) Assemble current viewing catalog (step 410).
[0176] a) Allocate a block of memory 120 which is sufficiently
large to accommodate viewing catalog 452 for the current week.
[0177] b) For each household in household object array 454, locate
the entry for the household in the viewing catalog for the previous
week.
[0178] i) If an entry exists in the viewing catalog for the
previous week, then make an entry for the household in the same
position in the viewing catalog for the current week.
[0179] ii) If an entry does not exist in the viewing catalog for
the previous week, then make a new entry for the household at the
end of the viewing catalog for the current week.
[0180] iii) Assign an entry in the viewing data 460 region of
memory 120 for the household. Indicate this entry position in
viewing catalog 452 for the current week.
[0181] c) For each household member in the person object array,
locate the entry for the person in the viewing catalog for the
previous week.
[0182] i) If an entry exists in the viewing catalog for the
previous week, then make an entry for the person in the same
position in the viewing catalog for the current week.
[0183] ii) If an entry does not exist in the viewing catalog for
the previous week, then make a new entry for the person at the end
of the viewing catalog for the current week.
[0184] iii) Assign an entry in the viewing data memory region for
the person. Indicate this entry position in the viewing catalog for
the current week.
[0185] d) Entries are not made in the viewing catalog for those who
are not members of households (visitors). Thus, for each person in
the person object array who is not a member of a household, assign
an entry in the viewing data memory region for the visitor.
Indicate this entry position in the person object.
[0186] 4) Read all program data records (both the lead and the
continuation records) from Nielsen data 430.
[0187] a) For each classification lead record, select the next
program object in the program object array for this program
data.
[0188] b) Assign data items in the program data lead record to the
program object selected for the program. Dates, times, and quarter
hour values in the program data item are converted from an ASCII
representation to a binary representation, and assigned to the
corresponding date/time attributes in the program object.
[0189] c) Increment the program object array counter in the Program
Group to indicate that the program has been added to the program
object array.
[0190] d) For each program data continuation record, locate the
Household or Person object referred to in the viewing event in the
Household Group or Person Group collections.
[0191] e) As described in the section "Viewing Data in the DME
Database", identify the memory location in the viewing data memory
for this household/person and date/time using the newly assembled
current week viewing catalog for households and household
members.
[0192] f) Enter in the viewing data memory location a notation
indicating that the household/person viewed the network program at
the indicated date/time.
[0193] 5) Read all usage data records (both the lead and the
continuation records).
[0194] a) For each usage lead record, identify the date and time of
the quarter hour viewing event for the lead record. This date/time
is used in reading each subsequent usage data continuation
record.
[0195] b) For each Household or Person in each usage continuation
record, locate the Household or Person object in the Household
Group 440 or Person Group 442 collections.
[0196] c) As described in the section "Viewing Data in the DME
Database", identify the memory location in the viewing data memory
for this household/person and date/time using newly assembled
current week viewing catalog 452 for households and household
members.
[0197] d) Enter in the viewing data memory location a notation
indicating that the household/person viewed the television at the
indicated date/time. This notation is conditional on the presence
of a preexisting notation:
[0198] i) If a notation has already been made in this location
based on a viewing event string in a Program Data Continuation
record indicating that the household or person was watching network
television, then no further notation is made. This assumes that the
usage viewing event refers to this network viewing event.
[0199] ii) If no notation is found, then it may be assumed that the
household or person was viewing non-network television, and a
notation is made accordingly.
[0200] 6) Write the allocated blocks of memory 120 onto disk in
binary form into a newly created .tvd database file 490 located in
DME database 126. In all cases, except as noted, the blocks of data
are appended contiguously to .tvd file 490.
[0201] a) Write header text which identifies the file for those
users who attempt to edit, type or print .tvd file 490.
[0202] b) Write a file format version number which can be read by
DME. 127
[0203] c) Write five long integers to .tvd file 490, each initially
have a value of zero. These values are used to note the offset
position in the database file of each of the arrays of objects and
data. Each of these values will be updated (as noted below) with an
actual offset value.
[0204] d) Write the viewing catalog 452 to .tvd file 490.
[0205] i) Write the length of viewing catalog 452
[0206] ii) Write the portion of viewing catalog 452 which has been
used. Note the offset position from the beginning of the file of
this catalog.
[0207] iii) Write the offset value to the first offset position as
described in (c) above.
[0208] e) Write the portion of household object array 454 to .tvd
file 490.
[0209] i) Write the number of entries in household object array
454.
[0210] ii) Write the portion of household object array 454 which
has been used. Note the offset position from the beginning of the
file of this catalog.
[0211] iii) Write the offset value to the second offset position as
described in (c) above.
[0212] f) Write the portion of the person object array 456 to .tvd
file 490.
[0213] i) Write the number of entries in the person object array
456.
[0214] ii) Write the portion of the person object array 456 which
has been used. Note the offset position from the beginning of the
file of this catalog.
[0215] iii) Write the offset value to the third offset position as
described in (c) above.
[0216] g) Write the portion of the program object array 458 to .tvd
file 490.
[0217] i) Write the number of entries in the program object array
458.
[0218] ii) Write the portion of the program object array 458 which
has been used. Note the offset position from the beginning of the
file of this catalog.
[0219] iii) Write the offset value to the forth offset position as
described in (c) above.
[0220] h) Write the portion of the viewing data 460 to the
file.
[0221] i) Write the number of entries in the viewing data.
[0222] ii) Note the offset position from the beginning of the file
of this data.
[0223] iii) Write the offset value to the fifth offset position as
described in (c) above.
[0224] iv) For each of the 28 periods of 6 hours each, write the
portion of the viewing data 460 which has been used.
[0225] 7) Close .tvd database file 490.
[0226] 8) Write current viewing catalog 452 to a separate file for
use in assembling the viewing catalog for the next week.
[0227] Although the example presented above specifically references
person-by-person media-related data for network broadcasting, cable
television network data can be added to the database file using
similar techniques. In addition, similar database files can be
created for other media types using similar techniques.
[0228] Sample Data Filtering Using DME Database
[0229] One significant advantage of the preferred embodiments for
DME database 126 is the very fast access to the data for purposes
of filtering data according to custom queries. Users will typically
want to analyze the viewing behavior of selected demographic groups
of people or households in the Nielsen sample for purposes of
analyzing behavior and targeting desired consumer groups for
advertising campaigns. The Nielsen data contains various data
elements that can be used for filtering. These elements include:
age; gender; income; level of education; profession; hours of
weekly television viewing; and the ages of family members that live
in the household. This data can be used to identify and select
various consumer groups for purposes of analysis relating to
advertising campaigns. For example, a user might wish to select all
women between the ages of 18 and 49 who live in households with
children, and having incomes greater than $40 K per year. Rather
than using complex mathematical relations to make this sample
selection, the data in DME database 126 is organized so that the
selection can be made using Boolean logic, which is relatively
fast, in order to compute using most typical computer systems.
[0230] Referring now to FIG. 5, the data elements for representing
the age of the sample audience members, for example, are not stored
as integers, but are stored as a 16 bit field where each bit
represents one of the available age ranges. The first several age
ranges are assigned bits in the field according to the table shown
in FIG. 5 Additional age ranges can be represented in a similar
manner. Given this structure, filtering the data contained in DME
database 126 for age-specific criteria is a simple Boolean
mathematical exercise. To select sample members based on age,
graphical user interface 125 utilizes a 16 bit age selection mask
with the required bits set to indicate the desired age range. If,
for example, the user wanted to analyze sample members in the 12-20
age range, then the age selection mask would be 8H+10H+20H=38H
(hex), or 0000000000111000 (binary). To determine whether or not a
particular sample member represented in DME database 126 is in the
desired age range, DME 127 performs a logical "and" operation using
the member's age field and the age-appropriate selection mask. In
the C programming language, this procedure may be represented as
"PersonAge & AgeSelectionMask."
[0231] If the person were actually in the age range of 15-17, then
the values would be 1 0000000000010000 ( age of audience member ) +
0000000000111000 _ ( age selection mask ) 0000000000010000
[0232] Which would result in a value of TRUE for the operation,
thereby indicating that the person is in the requested age
range.
[0233] Other similar operations can be envisioned to create filter
masks for other demographic information contained in DME database
126. Given this structure, a sample selection based on multiple
demographic fields would combine similar elements for those fields.
For example, in the C programming language, a selection based on
age, income, and education could be represented as shown
immediately below.
[0234] ((PersonAge & AgeSelectionMask) &&
[0235] (HouseholdIncome & HouseholdlncomeSelectionMask)
&&
[0236] (HouseholdEduc & HouseholdEducSelectionMask))
[0237] By providing the appropriate mask for the desired
demographic characteristics, DME database 126 can be quickly and
efficiently screened to locate the sample audience members who fit
the desired criteria. This kind of Boolean computation can be
executed very quickly on a digital computer. Alternatively,
although not preferred, a more conventional approach for a similar
selection might be the expression shown immediately below.
[0238] if (PersonAge>= LowerSelectionAge &&
[0239] PersonAge<= UpperSelectionAge &&
[0240] HouseholdIncome>= LowerSelectionIncome &&
[0241] HouseholdIncome<= UpperSelectionIncome &&
[0242] HouseholdEduc>= LowerSelectionEduc &&
[0243] HouseholdEduc<= UpperSelectionEduc)
[0244] The above expression is much more time consuming to evaluate
in a computer by virtue of the fact that it is longer and the math
operations (such as ">=") are more time consuming for processor
110 to evaluate than simple logical operations (such as "&")
are to evaluate.
[0245] Viewing Data in the DME Database
[0246] The Nielsen person-by-person data provides "viewing data"
for any given week. The viewing data indicates the viewing choices
made by sample households and members living in the sample
households for the midpoint of every 15 minute period during the
week. For example, for the week of Sep. 22, 1997, the data may
indicate that the viewing selection made by person number 2 in
household number 200011 at 8:08 PM on September 24; the midpoint of
the 8:00 PM to 8:15 PM quarter hour. The viewing options for this
person include at least three distinct options: 1) watching one of
the broadcast networks--ABC, CBS, Fox, or NBC; 2) watching
non-network programming such as unaffiliated stations or cable; or
3) turning the television off, i.e. not watching television. As
Nielsen makes other notational options available, such as including
the Warner Brothers Network, these new options can also be noted in
the data structure without modification.
[0247] If a person was not watching television during a given
quarter hour, then no viewing records are present in the Nielsen
data for that quarter hour. In addition, if the person was not
in-tab for that quarter hour, then the data delivered by Nielsen
indicates that condition as well.
[0248] In DME database 126, the sample viewing data provided by
Nielsen for a given week requires about 7 Mbytes of storage space.
In order to conserve memory space during subsequent processing and
analysis, it is desirable to avoid allocating memory for an entire
week of data when a user requires access to only some small portion
of it. Therefore, a week of Nielsen viewing data is divided into 28
blocks of about 250 Kbytes each, with each block representing six
viewing hours during the week for all households and people in the
sample. The broadcasting week begins at 6:00 AM on Monday morning.
Thus, the first block of viewing data begins at 6:08 AM on Monday
and ends at 11:53 AM. Similarly, the second block begins at 12:08
PM and ends at 5:53 PM, etc.
[0249] Each of these 28 blocks for a given week contains all of the
viewing data for a six-hour period either for all sample households
and members. Thus, if a system user requested viewing data for a
particular member of the sample at a specific time, then the
appropriate block of data in DME database 126 will be retrieved
from DME database 126 and loaded into memory 120. This block will
be the block that contains all the viewing data for a six-hour
period (including the requested time) for all members of the sample
audience. Alternatively, the broadcast week could have been
subdivided into a greater or lesser number of blocks by selecting
an alternative size for each of the blocks.
[0250] This memory management procedure is consistent with
anticipated mode of system use for a typical user. Typically, if a
user requests a type of analysis which requires viewing data for a
given sample member at a particular time, then the desired analysis
will generally also require viewing data for many or most other
members of the sample audience for the selected time and for
adjacent times. Thus, all necessary data is efficiently loaded as a
single block from DME database 126 into memory 120. This is a more
efficient process than would otherwise be required using
conventional database management systems which repeatedly return to
the database file for more data for other sample members or for
other times in an iterative fashion.
[0251] Referring now to FIG. 6, a simplified graphical
representation of the data contained in a .tvd file as stored in
DME database 126 is shown. The data in FIG. 6 represents a total of
7 households, with 17 members residing in those households, and 3
visitors. The indices in each of the arrays indicate a relationship
to the data contained in other arrays, and the viewing status
elements in the viewing data arrays are typical for actual members
of the sample audience. Although somewhat involved, FIG. 6 presents
a useful example of how the actual database arrays relate to each
other. In an effort to avoid too much confusion in the figure, not
all possible relationship arrows are included.
[0252] The arrows in FIG. 6 indicate some of these index
relationships between various data elements. Normal programming
practice in C++ suggests the use of memory pointers rather than
indices to relate one object to another. However, as explained
above, it is unknown where in memory 120 the allocated blocks will
be located after they are retrieved from DME database 126. So,
using the various preferred embodiments of the present invention,
pointers cannot be used to access database 126. The indices are
used in place of pointers to indicate an offset into each block of
data.
[0253] Each cell 651 in a block of viewing data 650 indicates a
television viewing status element for one member of the Nielsen
sample, or for one household, for the mid-point of one quarter hour
time period 651. If, for a particular record in the viewing data,
the person or household is watching one of the viewing options
(such as ABC, CBS, Fox, NBC, one of the cable networks, or
non-network programming), it is indicated. If the member was not
watching television, then cell 651 is blank. Also indicated is
whether or not the sample member is out-of-tab (shown in as an "O"
in each out-of-tab cell.)
[0254] DME database 126 includes a viewing catalog data structure
640 that relates person and household objects to viewing data 650.
The data for each household and person includes an index value 641
indicating the position in viewing catalog 640 for that person. So,
to retrieve viewing data for one member of the Nielsen sample for a
single quarter hour period, DME 127 will perform the following
tasks.
[0255] First, DME 127 will allocate a block of memory 120 for all
person objects for the desired week, and loads all person data
objects from disk into the allocated block of memory 120. Next, DME
127 allocates a block of memory 120 for all household objects for
the week, and loads all household data objects from disk into
memory 120. Then, DME 127 will allocate a block of memory 120 for
viewing data 650 and load the desired six-hour block of viewing
data 650 into this block of memory 120. Once viewing data 650 is
loaded into memory 120, DME 127 will allocate a small block of
memory 120 and loads viewing catalog 640 into this block of memory
120. Next, DME 127 will locate the person object in memory 120 for
the requested member of the sample and move to the position in the
viewing catalog as indicated by the catalog index value in the
person object. Then, DME 127 moves to the appropriate record in the
viewing data as indicated by the viewing index value in the viewing
catalog. Finally, DME 127 can move along the record in viewing data
650 to the desired time during the six hour time block and retrieve
the viewing indicator.
[0256] For example, to find out what person #4 in household #200143
was watching at 8:23 PM, DME 127 will search through the array of
person objects 630 until it finds person #4 in household #200143.
In searching for this particular person, the household number of
each person is found by reading the household index number, which,
in this case, is 6, and then reading the household number for array
element number 6 in the household objects array 610. After finding
the person, the catalog index number is read, which, in this case
is 27. Then, the 27.sup.th element of viewing catalog 640 is read
for the viewing index, which is 25. Next, the 25.sup.th element in
the person viewing data array is accessed. Finally, by referencing
the appropriate cell, this element indicates that at 8:23 PM the
person was watching non-network television.
[0257] There are at least two important reasons for redirecting the
viewing data locations through viewing catalog 640 rather than
indicating the location of viewing data with viewing data block 650
directly from household objects array 610 and person objects array
630. First of all, by redirecting the data lookup through viewing
catalog 640, viewing catalog 640 can remain consistent from one
week to the next. "Consistent" in this context means that the
various entries in viewing catalog 640 are in the same relative
position from week to week. The entry for person #1 in household
#200143, for example, is always six positions following the entry
in viewing catalog 640 for person #1 in household #2000013. Given a
consistent viewing catalog 640, it is only necessary to load a
single person objects array 630 or one household objects array 610
from a single week in order to retrieve data spanning multiple
weeks. The catalog index values in these arrays can be used with a
viewing catalog 640 from any week of data. Second, in order to
maintain consistency between weeks in viewing catalog 640, when
members drop out of the sample, empty space will remain in viewing
catalog 640. Viewing catalog 640 is not compressed to eliminate the
spaces, as indicated by the empty cells shown in viewing catalog
640 at positions 15, 16, and 19-20. But, because of the indirection
in the viewing index, these spaces are not necessary in the viewing
data, thus reducing the memory requirements for the data.
[0258] DME 127 can also be used to review the viewing habits of
person #1 in household #200143 over a period of several weeks. As
indicated earlier, for the week shown in FIG. 6, the catalog index
number for of person #1 in household #20014325, and the viewing
index number is 22. For the week shown in FIG. 6 viewing index
number 22 is used to retrieve the viewing information. However, in
order to analyze what this same person was watching during the
following week, there is no need to load another array of person
objects 630 and find the catalog index value for this person for
the next week, the existing index can be used.
[0259] The catalog index values remain the same for all members of
the sample from one week to the next. This person's catalog index
value is still 25. So, DME 127 loads viewing catalog 640 for the
next week and retrieves the viewing index value from catalog index
position 25. This viewing index value may not be 22 as it was in
the first week. If there is no viewing index value in viewing
catalog 640 at position 25 for the next week, it may be assumed
that this person was dropped from the sample, and that there is no
viewing data contained in viewing data block 650 for the person
during that week.
[0260] In summary, the use of viewing catalog 640 eliminates two
time and memory consuming tasks in retrieving viewing data which
spans days or weeks. First, there is no need to load person and
household objects for multiple weeks. Next, the need to search
through multiple person or household arrays for sample members is
also eliminated.
[0261] The above described architecture for DME database 126 is not
absolutely essential for the implementation of graphical user
interface 125. However, because of the significant speed advantages
afforded by this structure, it is currently the most preferred
embodiment for storing Nielsen data for use with the present
invention. An alternative preferred embodiment is described in
conjunction with FIG. 12. Future advances in computer hardware may
make it possible to implement the present invention using
conventional database management techniques. However, the specific
database designs of the present invention as described within this
specification will still provide a significant speed advantage over
other database structures presently known.
[0262] Referring now to FIG. 7, three viewing catalogs 640 for
three consecutive weeks are shown (week 1, week 2, and week 3). In
the week following that shown in FIG. 6, the first three households
(up through household #200045) are dropped from the sample. As
shown in FIG. 7, viewing catalog 740 has blank spaces in cells 0-7.
There is no viewing data for these households. Therefore, it is no
longer necessary to store cells 0-7 in the database, as indicated
by the dashed lines in FIG. 7. For week 2, viewing catalog 640 will
begin with the cell that corresponds to cell #8 of the previous
week, together with the absolute cell position of this first cell.
The number of the first valid cell in the viewing catalog is
referred to as the catalog offset (in this case, 8).
[0263] To extend the example, if households #200071 and #200102 are
dropped in week three, then the cells which correspond to cells
0-14 of week 1 will be blank. Recognizing that cells 15, and 16
were already blank, the first valid cell for week three is #17.
This is illustrated in FIG. 7 for viewing catalog 640 for week 3.
Similarly, only those cells of viewing catalog 640 beginning with
cell 17 along with this offset number are stored. From week to
week, the index to viewing data for each sample member is stored in
the corresponding cells of viewing catalog 640 as shown in FIG. 7.
For example, the catalog index for week 1 for person #3 in
household number 200143 is 26, for week 2 it is 18, and for week 3
it is 9. Note that the viewing catalog index values (the numbers
down the left hand side of each viewing catalogs 640) change from
week to week, but the relative positions of the cells do not.
[0264] Using the catalog index for any weeks, along with the
appropriate catalog offset values, DME 127 can compute the catalog
index value for any week in DME database 126. This capability
allows DME 127 to avoid needlessly searching through person object
arrays or household object arrays for other weeks that include the
sample member of interest. For example, it will be fairly simple to
retrieve the viewing data for person #3 in household number 200143
for all of the above weeks. Although any week could be used as the
starting point, for illustrative purposes week 2 is selected. In
week 2 the catalog index for person #3 in household number 200143
is 18.
[0265] The catalog index for any other week is a combination of
this catalog index and offsets for the two weeks, as calculated
below.
Catalog Index n= Catalog Index 0+ Offset 0-Offset n
[0266] Therefore, the catalog index value for week 1 is:
Catalog Index 1=18+8-0=26
[0267] and for week 3 the catalog index value is:
Catalog Index 3=18+8-17=9
[0268] The calculation of the catalog index value can be seen in
FIG. 8 Now, using a combination of the viewing catalogs 640 and
catalog offsets for several weeks, the viewing information for a
selected person in the Nielsen sample can be quickly accumulated.
First, the viewing catalog index number for the desired person is
retrieved from any one of the weeks of interest. Then, each of the
viewing catalogs 640, along with the associated catalog offset for
the catalog, is successively loaded for person #3 in household
number 200143, along with the viewing data for the associated week.
Then, by comparing the catalog offset, the viewing index, and
catalog values, one of the following conclusions will be reached.
If, for any given week, the calculated catalog index is less than
the catalog offset for that week, then the desired person has been
dropped from the sample. Additionally, if the location in the
viewing catalog for the catalog index plus the catalog offset is
blank, then the desired person has been dropped from the sample.
Finally, if the calculated catalog index is greater than the size
of the viewing catalog, then, for that specific week, the person
has not yet been added to the sample.
[0269] Most types of analysis dealing with advertising and
broadcasting require retrieval of media exposure data for
demographically related groups of people, not just individuals as
in the example above. For example, it might be desirable to know
what people in the 18-49 age group were watching at time 1 during
week 1, time 2 during week 2, and at time 3 during week 3. DME
database 126 is designed so as to be particularly well suited for
these types of queries. This is illustrated in the following
pseudo-code shown immediately below.
[0270] 1) load the person and household objects for week 1
[0271] 2) for each of three weeks
[0272] 3) load the viewing catalog
[0273] 4) load the catalog offset
[0274] 5) load the viewing data
[0275] 6) next
[0276] 7) for each person in week 1
[0277] 8) if this person is in the selected demographic group
[0278] 9) for each week n
[0279] 10) calculate the catalog index
[0280] 11) retrieve the viewing data for time n
[0281] 12) if this person is not in-tab or in the sample
[0282] 13) go to the next person
[0283] 14) end if
[0284] 15) next
[0285] 16) add results to summary values
[0286] 17) end if
[0287] 18) next
[0288] 19) return the summary values
[0289] Viewing catalog 640 will not grow indefinitely because
viewing catalog index positions are assigned to households and
people in the order in which they are added to the Nielsen sample.
Therefore, the sequence in which they are dropped from viewing
catalog 640 will be in approximate chronological order. The sample
members that are most likely to be dropped from the sample are at
the top of the catalog index because they have been in the sample
the longest.
USER INTERFACE
[0290] Graphical user interface 125 provides access to DME 127 and,
by extension, to DME database 126 via DME 127. User interface 125
provides an opportunity for a media planner to distribute
advertisements over time or space based on actual or anticipated
individual or collective advertising exposure. There are several
unique characteristics available in conjunction with user interface
125 that are especially advantageous for analyzing media-related
audience access information, such as the Nielsen data for
television viewing. Each of these specific features is explained
below.
[0291] Referring now to FIG. 8, a method 800 for using a preferred
embodiment of the present invention to access the television
viewing data is described. System users can gain insight into how
audiences make television viewing decisions by using the system to
interactively browse through the viewing data. To use the system, a
person typically iterates through the steps illustrated in FIG. 8.
The user formulates a question or hypothesis about audience viewing
behavior (step 810). The user composes a query based on the
question or hypothesis using graphical tools supplied user
interface 125 (step 820). The user submits the query to the DBMS
(step 830). DME 127 selects a subset of the audience sample based
on demographic choices the user made in composing the query (step
840). DME 127 computes/tabulates the results (step 850) and returns
the results to system 100 (step 860). Then, user interface 125 of
system 100 presents the query results in graphical and/or tabular
form to the user (step 870). The user then examines the results,
and in doing so, may formulate new questions or hypotheses about
viewing patterns and decisions (step 890). In this case, and based
on these new hypotheses, the user may return to step 810 as often
as desired in order to compose one or more new queries.
[0292] Referring now to FIG. 9, sample cross tabulation information
for the hit television program "Friends" is shown. This type of
graphical presentation for media-related data is not readily
available for general use in the market today. Typically, this type
of information is only available by contacting organizations that
specialize in producing it. However, with the various preferred
embodiments of the present invention, this type of information can
be made readily available to a large audience. Another feature of
graphic user interface 125 is the ease of selecting desired
demographic information. Demographic groups can be selected by
adjusting the length and position of a series of graphical bars, in
which the position of each bar represents the selected range for a
single demographic attribute. The user clicks on the numerical
values indicating the selected range of values. The bar position is
adjusted to reflect this selection. In addition, a user of system
100 can immediately access a variety of useful media-related
person-by-person information by merely clicking on a single
icon.
[0293] Similarly, referring now to FIG. 10, a user may "click" a
mouse on the defection icon and generate the line graph shown in
FIG. 11. The icon-driven graphical user interface 125 provides
single click access to very sophisticated types of information.
Anywhere on any screen where a program names or data is displayed,
the user may retrieve more detailed information on a given program
by selecting the program name or data region using the mouse.
Alternatively, if a user selects a program as described above, the
system could be configured to display historical ratings trends.
Finally, the user can customize the system to determine the
information that is displayed when the user selects an item. Some
useful items that can be accessed via user interface 125 are
explained briefly below.
[0294] Program Lists. Using a variety of interactive mechanisms, a
user can assemble lists of program episodes. Analysis can then be
performed on these lists.
[0295] Program Schedule Data Dynamics. A user can select a program
plan or schedule for display, and then select other data elements
for display in the context of the plan or schedule. The user can,
for example, select for display the programming schedule for NBC
for all Monday evenings between two dates. The user then could
select for display adjacent to the name of each of the programs the
retention or lead-in value for the program.
[0296] Advertising Exposure Valuations
[0297] Some of the features of graphical user interface 125 provide
mechanisms allowing the user the ability to assign advertising
response values to various selected alternatives. This allows a
media planner to perform "what-if" analysis to compare various
options and determine which options are most viable. In addition,
costs for various advertising exposure options can be assigned
based on time or space boundaries for the purpose of scoring or
valuing various alternative options for an advertising plan. For
example, the media planner can graphically interact with the
mechanisms of user interface 125 to select various options from a
variety of alternatives, thereby arranging a proposed or actual
advertising schedule in space and time. User interface 125 will
provide real-time feedback for comparing the various options as the
media planner cycles through the available choices to determine the
most effective use of resources. Through the mechanisms of user
interface 125, the media planner can specify "space" boundaries for
a given advertisement or group of advertisements, thereby
maintaining a specified distance from other advertisements. The
mechanisms of user interface 125 can also provide information
regarding the estimated influence of advertising messages on
individuals or audiences base on many factors such as exposure
influence over time based on the declining influence of advertising
over time, the accumulated influence effect of multiple exposures
over time, influence due to frequency of exposure to the
advertisements, etc.
[0298] A computer system 100 for data manipulation and analysis in
accordance with a preferred embodiment of the present invention
employs a unique user interface 125 which, in conjunction with DME
127, can retrieve the Nielsen data from DME database 126 and then
present the data in graphical and tabular forms to system users.
Then, combining this information in various ways using advertising
optimization mechanism 128, advertizing decisions can be optimized
for practically any desired set of objectives. The various
embodiments of user interface 125 are designed to be easy to use
and intuitively simple. This allows broadcasting and advertising
professionals to understand the viewing patterns of the television
audience with little or no formal training and to quickly and
easily arrive at optimal advertising solutions for the desired
objectives.
[0299] System users are often interested in the television viewing
behavior of particular demographic groups. They may, for example,
be interested only in adults in the age range of 18 to 49 years old
who live in the northeast United States, and who live in households
with incomes greater than $40,000 per year. The data analysis
system of the present invention is designed to provide convenient
isolation of these types of demographic groups in the sample, and
the necessary tools for analyzing their viewing habits.
ADVERTISING OPTIMIZATION MECHANISM
[0300] 1. Introduction
[0301] A significant challenge in media advertising is to identify
optimal compromises in satisfying simultaneous but sometimes
conflicting advertising objectives. The media advertising planner,
for example, would like to advertise only to those people who
frequently use an advertised product or service, but he recognized
that many or most people, to some degree, may be potential product
or service users; that an advertising audience is composed of a
range potential for using an advertised product or service. The
planner targets one well-defined demographic group in terms of age,
gender, income, education, territory, etc. but also recognizes that
people in other demographic groups may also have some marginal
value as potential customers. The planner wants broad advertising
reach, but only to the extent that it is cost effective and does
not result in excess exposure. The audience needs to be informed
about the product but not saturated with the advertising message.
In addition, advertisements need to be aired on specific days or
times so that the message is still fresh in the minds of the
audience when they are ready to purchase. Finally, the
advertisements also need to be placed on programs where the
audience is attentive, and where the programming is consistent with
the advertising message.
[0302] The range and volume of data available to a media planner
that can be used to develop an advertising plan can be formidable.
Data can be readily licensed from a variety of sources that detail
a variety of elements important to the planning process. Some data
measures individual and aggregate exposure to advertising media
using a breadth of demographic and other factors. Other data
specify the product and service usage habits by either demographic
group or media usage. Media planners readily agree that the data is
potentially valuable in assessing the merits of a media plan. The
challenge however, is to systematically use all of these data in
guiding the development of a optimal plan. All of the data is
systematically considered and weighted at the same time.
[0303] The following portion of this specification describes the
preferred embodiments of an integrated method for optimizing the
scheduling or positioning of advertisements and promotions in a
media environment. The integrated method is accomplished by using
advertising optimization mechanism 129 (shown in FIG. 1). The
method is integrated in the sense that it considers a comprehensive
set of factors for identifying optimal plans or schedule, including
product or service usage, reach, frequency, learning, timing,
demographics, viewer response, and cost. Using this method, all of
these factors can be considered simultaneously in the decision
making process. This is accomplished by measuring the achievement
of specific objectives based on these factors using detailed
historical audience exposure data, and then merging the individual
factor measurements to arrive at a comprehensive indicator of
advertising success.
[0304] The various preferred embodiments of the optimization
methods described herein may be used at several different points
during the process of developing a comprehensive advertising
campaign. It could be used early in the planning process to test
the sensitivity of selected advertising objectives against media
vehicles. Later in the process, it could be used to build plans or
schedule, or test modifications to a plan or schedule when more is
known about the availability of advertising slots. Finally, after
an advertising campaign begins, the system could be used to compare
planned versus actual objectives, to monitor the effectiveness of
the campaign, and to adjust the plan or schedule to make up for
deficiencies. In cases where there is detailed information
available to the planner about market conditions and consumer
characteristics, the method consistently and systematically applies
this information in the decision making process. When very little
information is available, the method can still be used by the media
planner in investigating the efficiency of past advertising
campaigns, and in improving the planner's understanding of audience
exposure patterns.
[0305] In addition, the various preferred embodiments of the
present invention can significantly improve the efficiency of
creating advertising plans or schedule for a variety of media types
(i.e., print media, radio, television, etc.). The methods of
optimization presented herein consistently include much more
relevant data than many other planning techniques generally in use
today. The methods embodiment in advertising optimization mechanism
129 provide the framework for describing all types of media
objectives and presents a knowledgeable user with a complete
complement of tools necessary for quickly and easily performing
what-if studies and making profitable advertising decisions.
[0306] 2. Detailed Description
[0307] Advertising Optimization
[0308] The process of optimizing an advertising plan or schedule
according the preferred embodiments of the present invention is an
incremental one. It begins in the same way that conventional media
planning processes begin, by defining a set of media objectives
that are generally based on market objectives and research
information.
[0309] Referring now to FIG. 13, a process 1300 for optimizing an
advertising plan or schedule according to a preferred embodiment of
the present invention is described. Typically, an initial based
plan or schedule 1310 is prepared, again, in generally the same way
that advertising plans or schedule are prepared currently. However,
after the base plan or schedule is prepared, adjustments,
additions, and deletions are made in a very different manner. This
is where optimization process 1300 takes over, by incrementally
modifying the base plan or schedule to more closely meet the set of
media objectives, or to reduce cost. Several spots are selected by
the media planner as possible additions to (or deletions from) the
plan or schedule (step 1352). Each of these alternative spots is
scored according to its ability to efficiently contribute to
meeting the objectives (step 1358), resulting in a series of scores
1360. It should be noted that historical viewing data 1312, market,
program and audience research 1314, market objectives 1316, media
objectives 1320 are all possible factors to be used in the scoring
process. In addition, viewing data for an optimized plan or
schedule 1380 may also be used as a feedback mechanism for
computing scores.
[0310] Then, based on these scores, the media planner selects one
of the alternative spots to add to (or delete from) the advertising
plan or schedule. This improved plan or schedule 1364 then becomes
the base plan or schedule upon which further modifications can be
made using this same optimization technique (step 1362). The media
planner continues to iterate through this process until satisfied
with the optimized plan or schedule. At that point the final
optimal plan or schedule 1370 is achieved, the initial optimization
process ends, and the plan or schedule is executed (step 1375).
However, as shown in FIG. 13, part of the process may include
additional feedback and scoring, as desired.
[0311] Given this type of interaction, system 100 provides a user
interface 125 for optimization purposes and the necessary tools
(i.e. DME 127) to manipulate DME database 126. This allows system
100 to be used to program the plan or schedule and to rapidly
achieve the desired results. According to a preferred embodiment of
the present invention, an advertising plan or schedule is optimized
one spot at a time. This is because the contributed value of adding
an advertisement to a plan or schedule depends on what spots are
already in the plan or schedule. In the case of broadcast
advertising, for example, an advertisement in isolation may have
one value or score, while the value or score of the same
advertisement could be far different if it were aired immediately
following another advertisement in the same advertising plan or
schedule.
[0312] As shown in FIG. 13, an optimization mechanism or process
1350 could be an interactive one between a media planner and system
100. Although scoring is not particularly complex mathematically,
it is data and computation intensive, and requires access to large,
person-by-person, media exposure databases, such as the Nielsen
data. In order to optimize an advertising plan or schedule, a media
planner should take several steps. First, the media planner should
use market, program and audience research input 1314 coupled with
market objectives input 1316 to define the desired media objectives
(step 1318). If, for any reason, an interactive approach is not
desired, the system can be configured to automatically iterate
through a series of options to find the highest scores.
[0313] In initiating the planning process for an advertising
campaign, media planners will typically select objectives for the
campaign. Currently, these objectives are often established in
terms of desired reach and frequency, and are based on experience
and techniques separate from the optimization techniques described
herein. One objective for an advertising campaign, for example,
might be to reach 60% of the adults 18-49 in a population with at
least three "opportunities to see" over a four week period. Based
on these objectives, then, advertising vehicles and advertising
plans or schedule are selected.
[0314] The preferred embodiments of the present invention described
herein expand on the options available to media planners for
setting advertising campaign objectives. The methods of the present
invention increase the flexibility available to media planners in
specifying reach and frequency objectives. For example, various
preferred embodiments of the present invention allow a media
planner to specify the relative value of several levels of exposure
at the same time. The media planner is no longer confined to
selecting a single level at which advertising exposure becomes
effective. Further, the present invention provides options for
specifying campaign objectives not previously used.
[0315] Media planners, for example, have long believed that
advertising exposure becomes increasingly valuable as the time
between exposure and the purchasing decision narrows. Without the
methods of the present invention, however, there has been no
systematic way of incorporating these beliefs into the planning
process. Therefore, the present invention not only outlines a
method which media planners can use to specify temporal advertising
objectives, but it also includes a method for using these
objectives in scoring specific advertising campaigns for
decision-making purposes. Using at least one preferred embodiment
of the present invention, for example, a media planner can specify
that an exposure on the same day as a purchasing decision for a
particular item has twice the value as an exposure on the day
previous to the decision.
[0316] Once identified, media objectives input 1320 become a
critical input parameter for optimization process 1350. In addition
to media objectives input 1320, historical viewing data input 1312
and the initial base plan or schedule input 1310 are used as
contributing factors for scoring various advertising choices in a
subsequent step. Historical viewing data input 1312 is typically
Nielsen person-by-person data or some other relevant source of data
regarding viewers and their person-by-person viewing choices.
[0317] Next, the media planner will identify and select (step 1352)
alternative spots input factor 1356 to add to or delete from the
initial base plan or schedule input 1310 to arrive at a new base
plan or schedule input 1354. Once new base plan or schedule input
1354 is in place, the media planner can start the computer-based
process of scoring each of the alternatives (step 1358). Step 1358
represents an automatic scoring process, based on predetermined
parameters, which is performed by system 100. The actual scoring
methodology is explained below in conjunction with FIG. 14. The
computed scores input 1360 is used to modify base plan or schedule
input 1354 (step 1362). From this point, the media planner can make
various decisions 1322 and return to step 1352 to analyze the value
of the selected changes. This iterative process continues until the
media planner arrives at the final optimized plan or schedule 1370.
Finally, the media planner will execute the optimized plan or
schedule (step 1375). Note that the media planner can have input to
modify the base plan or schedule each time through the loop by
making selections from among the scored alternatives. This input is
illustrated by media planner's decisions 1322.
[0318] So, as illustrated by the example above, method 1300 for
optimizing an overall advertising plan or schedule is a gradual
process. It is also important to note that, even after a plan or
schedule begins to run, a media planner may wish to review the
results of the optimized plan or schedule using actual exposure
data for the first few weeks of the plan or schedule. The feedback
loop in FIG. 13 indicates that viewing data input 1380 for the
optimized plan or schedule can become input to the process of
computing plan or schedule scores. This allows a media planner to
further refine an advertising campaign after implementation based
on results of the previously implemented decisions. This feedback
activity is very valuable to a media planner and represents yet
another way that system 100 can be used.
[0319] Scoring Methodology
[0320] It should be clear from the discussion above that the
process used to compute the scores of alternative spots is the key
to effectively and efficiently optimizing an advertising plan or
schedule. If the scoring process is flexible, accurate, and fast,
then a media planner should be able to search through many
alternative plans or schedule for one that is efficient in terms of
meeting the pre-identified media objectives, and, at the same time,
minimizes cost. This most preferred embodiments of the scoring
process implemented by system 100 is detailed below.
[0321] Referring now to FIG. 14, a scoring method 1400 according to
a preferred embodiment of the present invention is illustrated.
Scoring method 1400 may be used as step 1358 of FIG. 13. As shown
in FIG. 14, the total score for a given alternative spot is
preferably composed of five distinct measures or indices which are
combined into a single score. Each of these five factors is
computed based on media objectives and/or historical exposure data
for individual members of a sample audience. It is important to
note that in the most preferred embodiments of the present
invention, every element that could influence the value of an
advertising plan or schedule is included in one and only one of the
five factors. It is also important to note that certain of these
factors may be omitted. However, the most preferred embodiments of
the present invention uses all five indices or factors.
[0322] This flexible scoring design allows a media planner to use
all five of the indices, if desired. Alternatively, a media planner
may wish to generate an optimum advertising plan or schedule using
limited data that does not need input data for all five indices. If
the relevant data regarding host media data required to generate a
response index are not available, the media planner can ignore the
response index. The resulting optimum campaign may not reflect the
influence of variations in response, but it may be adequate for
early planning purposes. Similarly, if a media planner chose to
ignore advertising spot timing while, at the same time, factor in
detailed demographic data, this could also be accomplished.
[0323] The computed value for each of these five factors is
referred to as an index, and as such is an indicator of relative,
rather than absolute, value. The audience valuation index, for
example, is a number that indicates the relative value of audience
members to a particular advertiser. The value has no units of
measure, such as $/person, but is expressed as a percentage. If one
plan or schedule has an exposure valuation index of 200% for an
advertiser, and another is rated at 100%, then the first has twice
the value to that advertiser as the second does. This allows for
easy and accurate comparisons to be made.
[0324] In order to more accurately describe the scoring process and
methodology, the following section presents a brief overview of the
scoring process. First, each of the indices is described. Next, the
scoring process is presented and described in equation form.
Finally, a specific detailed example of how system 100 uses a
simple plan or schedule and a small audience sample to compute a
score.
[0325] Scoring Factor Description
[0326] According to a preferred embodiment of the present
invention, there are five indices used to score each alternative
slot or scheduling choice that a media planner may consider in the
advertising optimization process. These five indices are: an
exposure valuation index 1440; an audience valuation index 1450; an
exposure recency index 1460; a response index 1470; and a cost
index 1480. Each of these indices is explained briefly below. While
the descriptions below provide insight into specific preferred
embodiments, many variations are possible, based on the preferences
and goals of the media planners implementing the scoring
methodology.
[0327] Exposure Valuation Index 1440. Exposure valuation index 1440
is the sum of the value of all individual audience exposures to an
alternative spot. It reflects that belief that the value of a first
exposure for an individual may not have the same value as a second
exposure, or a third exposure. Exposure valuation index 1440
represents a total exposure valuation, not an average. If an
alternative spot has many exposures because of scheduling or
vehicle popularity (i.e., the spot is shown during a highly rated
program), then the value of the spot is higher than if few viewers
are exposed to the spot. The value of each individual exposure is
computed using one of several techniques, which are all based on
two significant elements. Those two elements are exposure
objectives in the media plan, and the exposure history for the
individual audience member.
[0328] One technique, for example, depends on an assigned exposure
value for each frequency level. The first exposure to an
advertisement by an individual may be assigned a value of 1.0. A
second exposure may be valued at 2.5, etc. Then, the viewing
history for each member of an audience who is exposed to the
advertisement of interest is examined to determine the number of
times each has seen the advertisement previously. Finally, the
exposure value is returned for that frequency level. These exposure
values are summed across all exposures to produce exposure
valuation index 1440.
[0329] Suppose two people are together in a room watching
television when an advertisement is aired. The first person has
never seen the ad before, while the second person has already seen
the ad 100 times. The value to the advertiser of airing the ad is
different for the two people. The first person might learn
something about the advertised product, while the second person
will probably only be annoyed by having to view the ad for the
101.sup.st time. The value to the advertiser of exposing the first
person to the ad might be quantified as having a value of 1.0,
while the second person has a value of zero or less than zero.
Using a similar approach for the rest of the viewing audience, and
by summing the values of all of the individual exposures, the total
exposure value for this advertisement can be determined. This
number is a total exposure valuation, not an average. If an
alternative spot results in many impressions because of scheduling
or vehicle popularity, then the value of the spot is higher than if
few impressions result.
[0330] Note that the potential exposure value of the advertising
slot will be different for each advertiser and for each advertising
message depending on how frequently the ad has been aired
previously, when it was aired, and who in the audience has already
seen the ad.
[0331] In computing exposure valuation for specific dates and times
in the future, there are at least two possible sources of
respondent level data available for use in the evaluation. The
first is to identify a time in the past for which respondent level
data is available, during which time the media and advertising
conditions are similar to those anticipated during the advertising
campaign. Some of the considerations that might influence the
selection of the historical period include proximity to holiday
periods, sporting events, other seasonal conditions, etc. Exposure
valuation can be based directly on this historical data. The most
significant advantage to this approach is simplicity.
[0332] The second option would be to forecast viewing behavior at
the individual respondent level based on historical data. This
forecasted viewing could then be used as the basis for computing
exposure valuation.
[0333] Audience Valuation Index 1450. The target audience for many
media plans is narrowly defined using just a few demographic
parameters, such as age and gender; women 18-49, for example. The
demographic profile for the users of most products, however, can
usually be characterized more continuously using more dimensions.
Some audience age ranges, for example, may be more valuable to
particular advertisers than others, but many age ranges could have
some value. Similarly, other demographic parameters, such as income
and education, could also be used to characterize a typical product
user.
[0334] The total value of an audience member to an advertiser can
be computed using values assigned by the advertiser to the various
demographic characteristics. Each income level, for example, could
have an assigned value. These values for each of the demographic
measurements for each audience member are then combined to arrive
at a total value for that audience member.
[0335] Audience valuation index 1450, then, is a sum of the value
of individual member of the audience who is exposed to an
advertising spot. This is a total audience value, not an average.
If many people in a sample audience are exposed to an alternative
spot, then the audience valuation index for that spot will be
higher than if few saw the spot. The values assigned to the
individual members of the viewing audience are based on demographic
objectives in the media plan and these values will vary based on
the goods or services being advertised. The plan specifies a set of
values for each demographic characteristic. Each income level, for
example, could have an assigned value. Then, for those who were
exposed to the alternative spot, the assigned values are returned
for each demographic characteristic of interest. These values are
multiplied together to arrive at a total value for that audience
member. These values are then summed across all exposures, to
produce the audience valuation index. Again, this is a total
audience value, not an average. If many people in a sample audience
are exposed to an alternative spot, then the audience valuation
index for that spot will be higher than if few members in the
sample audience saw the spot.
[0336] Exposure Recency Index 1460. Exposure recency index 1460 is
an indicator of the timeliness of a spot based on advertiser
preferences for time of day, or day of the week. The purchasing
decisions for some types of advertised products are frequently made
at predictable times, and the media plan or schedule may indicate
the relative value of advertisement placement based on timing. The
influence that advertising exposure has on persuading people to
purchase the advertised products gradually declines over time, and
the more time that elapses between the time of exposure and the
time of purchase, the less influence the advertisement will have on
the purchase decision.
[0337] Assumed an advertiser has two identical advertising slots
with identical audiences available, one early in the week and one
on Thursday or Friday, the use of exposure recency index 1460 can
be demonstrated. If the advertiser knows that 30% of the purchases
of the advertised product occurs on Saturdays, then he will
probably consider the Thursday/Friday night advertising slot to be
more valuable than an earlier one. If for example, an early week ad
is worth $X to the advertiser, then a Thursday ad slot might be
worth $1.3X. The exposure recency index for the Thursday ad slot
would then be 1.30.
[0338] Response Index 1470. Response index 1470 is an indicator of
the average level of response that audience members are expected to
have as a result of being exposed to a given advertisement. This
value is probably judgmental in nature, and is dependent on a
number of factors. For network television, these factors might
include a variety of factors. For example, two important factors to
consider in response index 1470 might be average program attention
level for the program in which the advertisement is placed and the
consistency between programming themes and the advertised product.
Again, this highlights the correlation between the advertised
product or service and the target market.
[0339] It is important to note that response index 1470 should not
be dependent on factors which are accounted for in other indices,
such as program loyalty levels for series programs because loyalty
levels are related to frequency. Similarly, factors that reflect
audience skew toward one demographic group or another are not
included in the response index. As mentioned earlier, these
examples are merely representative, and should not be considered
exclusive or exhaustive. Obviously, those skilled in the art will
recognize that many other conditions, including media selection,
audience characteristics, and scheduling, may influence the
response indices assigned to specific advertisement alternatives.
These various indices will typically be specifically selected to
shift the scoring emphasis as desired for a given set of media
objectives.
[0340] In quantitative terms, if a program is judged to have a
response index of 1.0 for a particular product, and another program
has a response index of 0.5 for this same product, then the first
program, all other things being equal, should have twice the appeal
to advertisers of the product.
[0341] Cost Index 1480. Obviously, all other factors being equal,
as the total cost of a given spot increases, the attractiveness of
the spot to an advertiser declines. It is possible, however, that
based on the other four factors introduced in this section, an
inexpensive spot might not be worth the cost to a particular
advertiser while a relatively expensive one might be. Conversely,
some advertisers might be committing to relatively expensive spots
that have less value than more plentiful and less expensive
alternatives. The optimization techniques of the present invention
provide a comprehensive method for making this kind of quantitative
determination. Cost index 1480 simply tracks the absolute cost of
the alternative spot as measured in dollars. An important part of
any advertising campaign is to determine the appropriate tradeoff
between maximum desired exposure and finite constraints on
advertising dollars. While buying all of the available advertising
time on the Super Bowl will guarantee very broad exposure, most
media planners try to get the best "bang for the buck." Cost index
1480 will bring the dollar factor into the equation.
[0342] Scoring Equation
[0343] Symbolically, the computed score for an incremental change
in an advertising plan or schedule based on these five factors is:
2 S b ( a ) = i = 1 N a [ V l n ( i ) .times. d = 1 D V A d ( i ) ]
.times. V T ( a ) .times. V R ( a ) V C ( a )
[0344] where
[0345] S.sub.b(a) Total score for alternative spot a in plan or
schedule b.
[0346] N.sub.a Total number of exposures of spot a by a sample
audience.
[0347] D Number of demographic factors considered. 3 V d a ( i
)
[0348] The index value attributed to demographic factor d for the
audience member who saw exposure i out of N exposures to spot a. 4
d = 1 D V A d ( i )
[0349] The product of the index values of each of the D demographic
factors for one audience member. In other words, this is the total
indexed value for the demographic characteristics of a single
audience member to a particular advertiser for a particular
advertised product. 5 V I n ( i )
[0350] The index value of exposure i out of a total of N exposures
to spot a, which, for this member of the sample audience, is the
n.sup.th exposure to advertisements in plan or schedule b. 6 V I n
( i ) .times. d = 1 D V A d ( i )
[0351] The indexed value of a specific exposure times the indexed
value of the demographic characteristics of a single audience
member to a particular advertiser for a particular advertised
product. 7 i = 1 N a [ V I n ( i ) .times. d = 1 D V A d ( i )
]
[0352] The total value of exposing all members of an audience (or
all members of a sample group) to a particular advertisement at a
specific point in time. 8 V T ( a ) V R ( a ) V C ( a )
[0353] The response index value for spot a.
[0354] The cost index value for spot a.
[0355] Although the scoring equation is rather formidable looking,
a simple example is useful to show how the scoring process may be
accomplished by adding a particular advertisement to an advertising
plan or schedule.
[0356] Scoring Example
[0357] In order to further explain the preferred embodiment of the
scoring methodology of the present invention, equation (1) above
will be used to illustrate planning a new advertising plan or
schedule for network television. To simplify things, assume a total
audience sample of only 10 people. In the hypothetical planning
process, one new spot, spot D, is being considered for addition to
an advertising plan or schedule. The present advertising plan or
schedule has just three advertisements: spot A, spot B, and spot C.
To begin, a period of time during the previous month which has
programming similar to the period which the new plan or schedule
will span is identified. That period is used as the basis for
planning the future plan or schedule. The effect of adding spot D
to the plan or schedule will be examined in the context of the five
indices described above. Spot D represents an additional airing of
the same advertisement represented in spots A, B, and C.
[0358] Referring now to FIGS. 15-18, the operation of the scoring
methodology of the present invention is illustrated. FIG. 15
presents the information used to calculate exposure valuation index
1440. Exposure to each of the four spots, including spot D, by the
10 members of the audience sample as shown in the table depicted in
FIG. 15. The members of the audience sample are numbered 1-10. The
letter "Y" in the block below the audience member number indicates
that the audience member was exposed to the spot. The scoring
methodology of the present invention will be used to compute the
score that results from adding spot D to the plan or schedule.
[0359] As illustrated in FIG. 15, person #1 saw the advertisement
when it aired in spot D. In addition, person #1 has already been
exposed to the advertisement one previous time, when the
advertisement aired at spot B. Based on media studies for the
product being advertised, the ideal frequency, or total number of
desired exposures during the life of the advertising campaign is
three. Using three exposures as the ideal number, relative
frequency values have been assigned to each exposure as shown in
FIG. 16.
[0360] FIG. 16 clearly illustrates that the exposure to spot D for
person #1 is not as valuable as it would have been had it been the
third exposure for person #1, as it was for person #4. Therefore,
based on market and product research, this second exposure of
person #1 has a value which is 80% of what it would have been had
it been the third exposure for person #1: 9 V I 2 ( i ) = 0.8
[0361] Similarly, this exposure value also applies to person #7 and
person #8 because spot D represented the second exposure to the
advertisement for them as well. However, in the case of person #6,
spot D represented the first and only exposure to the advertisement
so the value of 10 V I l ( i ) = 0.4applies to person #6. By
extension, spot D represents the third exposure for person #4, so
11 V I 3 ( i ) = 1.0 .
[0362] The values for other members of the audience sample have
been similarly calculated and are shown in FIG. 15. As shown in
FIG. 15, some members of the sample audience, such as person #2,
were not exposed to the advertisement when it aired in spot D.
Therefore, the exposure to spot D for person #2 does not contribute
to the exposure valuation index score for spot D. The scoring
methodology of the present invention sums over the number of
exposures, not over the number of audience members.
[0363] Audience Valuation Index 1450. This index is related to the
individual demographic characteristics of the sample audience which
will view spot D when it airs. For this particular example, the
media planner is interested in only two demographic
characteristics: age, and household income. Assume that the values
assigned to these demographic characteristics are summarized in the
tables shown in FIG. 17 and FIG. 18. FIG. 17 is the age range and
assigned value for female members of the audience sample.
Similarly, FIG. 18 depicts the income range and assigned value for
the audience sample. The various weights for the values shown in
FIG. 17 and FIG. 18 are assigned based on market research that
indicates which consumer is most likely to buy the advertised
product. The consumer most likely to purchase the advertised
product is a woman between the ages of 18 and 34 with a household
income of at least $36,000 per year.
[0364] For purposes of this example, person #1 is a woman of age
37, living in a household earning $26-$30 K per year. The values
for the demographic characteristics being considered for this
person are: 12 V A l ( i ) = 0.70
[0365] (Age)
[0366] and 13 V A 2 ( i ) = 0.85
[0367] (Income)
[0368] Other demographic values, such as household education, are
not specified, so those values are all assigned a default value of
1.0 Thus, for person #1, the expression in equation (1) above is:
14 V I n ( i ) .times. d = 1 D V A d ( i ) = 0.80 .times. 0.70
.times. 0.85 = 0.476
[0369] According to the assumed criteria, exposing person #1 to an
advertisement in spot D is 47.6% as effective as it would have been
under perfectly optimal conditions where a woman of age 18-34 in an
upper income household saw the ad exactly three times.
[0370] To fully evaluate the score for spot D, this process is
repeated for all 10 people in the sample and then the individual
scores are summed: 15 i = 1 N [ V I n ( i ) .times. d = 1 D V A d (
i ) ]
[0371] where N is the total number of impressions for the sample
audience for this advertisement (i.e., spot D), which, in this
case, is five. Note that audience members who do not see the
advertisement in spot D do not contribute to the total score. For
purposes of illustration, assume that the other four people who saw
the advertisement were, on average, the same as person #1, so the
sum of all five is: 16 i = 1 N .476 = 2.38
[0372] Exposure Recency Index 1460. In order to calculate this
index, assume that the product being advertised is a product that
is frequently purchased on the weekends, such as movie tickets, or
meals at restaurants. The media planner has discovered, through
market research, that the best time to advertise this product is on
Thursday evening, and that advertising on Wednesday, one day from
the ideal time, is about 85% as effective. Since proposed spot D
represents an advertising air time on a Wednesday night program, it
will have a lower value than the Thursday night program. Thus: 17 V
T ( a ) = 0.85
[0373] Response Index 1470. The relevancy of response index 1470 is
related to the host program where the proposed advertisement will
air. In this case, the host program for spot D is a particularly
popular one, perhaps Seinfeld or Friends, for example. In this
case, most media planners would believe that associating their
product with this particular host program will lead to a more
positive image for the product and, in turn, increased sales of the
product. Level of involvement is a term that refers to the
attentiveness that a viewer exhibits when viewing a given program.
The higher the level of viewer involvement with a program, the more
likely it is that the viewer will retain the information presented.
The media planner also believes that the level of involvement for
this program is higher than average and, therefore, believe that
this involvement will carry into the advertisement, thereby making
the advertisement more effective. For these various reasons, the
media planner has concluded that advertising on this program is
135% as effective as advertising on an average program. Thus:
V.sub.R(a)=1.35
[0374] Cost Index 1480. Finally, as would be expected, the cost of
advertising on this relatively more desirable program is higher
than average when compared to other, less desirable programs.
Therefore, the media planner would assign a relatively higher value
for cost index 1480 to this program. Thus, to include spot D in the
advertising plan or schedule for this program:
V.sub.C(a)=1.50
[0375] It should be noted that an entire list of relative cost
indices can be prepared for any given product, making the
application of including cost index 1480 relatively straight
forward.
[0376] Finally, given all of the scores for the various indices,
all of the pieces for computing the total score for spot D are
available and the indices can be combined (step 1490) by using the
previously explained equation: 18 S c ( a ) = i = 1 N [ V I n ( i )
.times. d = 1 D V A d ( i ) ] .times. V T ( a ) .times. V R ( a ) V
C ( a )
[0377] =2.38.times.0.85.times.1.35.div.1.50
[0378] =1.82
[0379] At this point the media planner can compare the score for
spot D against the scores of alternative spots E, F, and G which
are have been previously scored using the same criteria. The higher
the score, the more efficiently the selected advertising plan or
schedule will match the predetermined media objectives. Using this
comparative information, a media planner is in a position to make
an informed decision on how to best expand the simple advertising
plan or schedule from three spots to four spots. To expand the
campaign to five spots, the process can be repeated again. The
media planner selects a set of alternative spots, scores them and
then selects an alternative based on the scores. Although this
process can obviously be performed by hand, it would be an
extremely tedious and error-prone process. For this reason, the
most preferred embodiments of the scoring mechanism for the present
invention are implement in a computer-based optimization mechanism
as depicted in FIG. 1. Using a computer-based system, the process
of evaluating even a large plan or schedule requires only a few
minutes.
[0380] Although some of these above-mentioned factors are
occasionally used in media planning, the methods for computing and
combining these factors in ways that contribute to informed
decision making is unique. The advertising optimization methods of
the present invention are sufficiently flexible to accommodate a
wide variety of beliefs about advertising, how audiences behave,
and how they respond to advertising. The methods also provide a
variety of mechanisms for explicitly including virtually any type
of information that may contribute to or detract from the value of
advertising exposures. While there are some assumptions made along
the way, the assumptions are quantified and remain constant from
one evaluation to the next, assuring a consistent application of
the assumptions to the data. This will provide a more useful
relative index for purposes of comparison between various
advertising options. Finally, the advertising optimization methods
of the present invention provide a way for systematically and
consistently applying information and beliefs to the decision
making process so that the resulting decisions can be entirely
consistent with the information available.
[0381] Detailed Explanation of Scoring
[0382] Although the basic concepts of scoring have been explained
above, there are additional details that require a more significant
background for purposes of explanation. This section will provide
additional background information and specific detail on other
unique aspects of the various preferred embodiments of the present
invention.
[0383] Although advertisers pay standard rates for audience
exposures, measured in cost per thousand (CPM), exposures are not
of equal value to advertisers. The true value of an exposure is
based on a variety of factors beyond just age and gender, such as:
the number of times that an audience member has already seen the
advertisement; exposure spacing; other individual and household
characteristics; the time of day; media type; elements surrounding
the advertisement; time of year; nature of the product being
advertised; and buying habits of the individual. These factors may
include: the number of times that an audience member has already
seen the advertisement; exposure spacing; various individual and
household characteristics; the time of day; the type of media;
elements surrounding the advertisement; the time of year; and the
nature of the product being advertised.
[0384] All of these factors may combine and contribute to the value
of an audience member being exposed to a specific advertisement. In
fact, given sufficient data by which to compute the value of an
exposure, each instance in which a member of an audience is exposed
to a specific advertisement could have its own unique value to an
advertiser, and that value could be different for each advertiser.
An important objective for advertisers, then, is to plan or
schedule their advertisements so that many audience members that
they consider valuable will see the advertisement, while, at the
same time, avoiding plan or schedule choices and positions where
the proportion of valuable audience members is relatively low.
[0385] While this line of reasoning may, at first, appear obvious,
it includes a point of emphasis which is missing in many known
techniques. The focus of media planning using the preferred
embodiments of the present invention is on computing the value of
each individual exposure for each individual member of an audience,
not on simply estimating audience value based on demographic
groupings, or basic factors such as estimated reach and frequency.
Using the optimization mechanism of the present invention, the
value of an entire advertising plan or schedule can be computed by
computing the value of each individual exposure and then summing
these exposure values. The whole is the sum of the parts.
[0386] An optimization "objective function" as used herein is an
expression that is to be maximized in order to optimize an
advertising plan or schedule. Using the methods and techniques of
the preferred embodiments of the present invention, an objective
function is formulated from a set of factors that are derived from
media objectives. These factors, at the simplest level could
include familiar expressions for target audience age, gender, and
reach. At the other extreme, they could include elements describing
media characteristics, advertised product usage by audience
members, and exposure timing.
[0387] The most preferred optimization methods of the present
invention use five factors or categories of data. All conditions
that might influence the value of an advertising campaign falls
into one of the five categories. If a condition can be measured, in
can be included for consideration in the objective function and it
will thus influence the optimization process. As explained earlier,
these five factors are: multiple exposure value; audience value;
timing; response; and cost. Each of these five factors is explained
in more detail below.
[0388] Multiple Exposure Valuation
[0389] The multiple exposure valuation factor embodies the notion
that people will respond differently to an advertisement depending
on how many times they have already been exposed to it, and when
they were exposed. In the case of television advertisements, for
example, a room full of demographically identical people could all
see an advertisement at the same time, but have dramatically
different reactions to the advertisement based solely on their
individual exposure history for the ad.
[0390] For example, a person who sees the advertisement for the
first time may not fully understand what is being advertised. With
a second exposure, a person may listen more attentively, or, having
seen it previously, may ignore it. Further, the third exposure to
the same advertisement may convey enough information to motivate
the person to actually try the product. Alternatively, if it has
been many days or weeks since the previous exposure, much about the
advertised product may have been forgotten, and an additional
exposure may have the same influence as the first exposure did.
Finally, if a person has already had many recent exposures to a
given advertisement, then a subsequent, new exposure may be ignored
entirely. The value of a single exposure, then, can be determined
only in the context of other exposures in an advertising plan or
schedule. One cannot place a value on an advertising exposure for
an individual without knowing what other exposures the individual
has had. The first exposure could be quite valuable, but the
twentieth exposure to a given advertisement during a given week may
have no value.
[0391] Quantifying the multiple exposure valuation factor, then, is
the process of estimating the value of an individual exposure based
on the position and timing of other exposures in the same
advertising plan or schedule. Many studies have been conducted,
concerning the influence that multiple advertising exposures have
on subsequent purchasing behavior. Researchers have been interested
in a number of issues, such as how many exposures are required to
convert an audience member into a purchaser, the ideal spacing in
time of advertising exposures, and message saturation that might
occur after being exposed to many advertisements. The methods and
techniques of the present invention as described herein make no
rigid assumptions about these issues. It is likely that a variety
of factors, including such things as product type and advertising
vehicle, will heavily influence the audience response to multiple
advertising exposures. The preferred embodiments of the present
invention do, however, provide a flexible framework by which a user
can specify the value of multiple exposures to a particular
advertising message. The system then optimizes a plan or schedule
based on those predetermined specifications.
[0392] The techniques used in formulating an advertising campaign
are drawn from the media objectives that a media planner already
has available. These techniques can range from the relatively
simple to the complex. The simple techniques are easy to use, and
require very little data, but may not fully describe the influence
of multiple advertising exposures on particular audiences. The more
complex methods require more data and processing, but the resulting
plans or schedule will be more efficient and consistent with the
media plan, and with assumptions about how people respond to
advertising.
[0393] For the purposes of the present invention, an efficient
advertising plan or schedule is defined as one that exposes an
audience to an advertising message in a way which is consistent
with a predetermined set of media objectives, and one which does so
at the least cost. A very simple set of media objectives, for
example, might include requirements such as exposing women 18-49 to
a given advertisement at least three times over a four-week period.
Given this objective, and in comparison with less efficient plans
or schedule, an efficient advertising plan or schedule would do any
one of the following: have fewer women in the age range who are
exposed fewer than three times for the same cost; have fewer women
in the age range who are exposed more than three times at a reduced
cost; have more evenly distributed exposures throughout the four
week period for more of the audience for the same cost; expose
women equally at a reduced cost. While all of these elements cannot
be satisfied simultaneously, the purpose of the scoring system is
to consistently weight each of these elements (and other elements
which will be introduced shortly) to arrive at an optimum
advertising plan or schedule. At least five general techniques can
be used to estimate the value of multiple advertising exposures.
The explanations presented below will begin with the simplest
technique and progress to the more complex. These various
techniques are summarized in the table shown in FIG. 19
[0394] Average Frequency Technique
[0395] Average frequency is defined as the average number of times
an audience is exposed to an advertising vehicle over a given
period of time. The time span is sometimes referred to as the
purchase cycle for the product advertised, and is, by convention,
often four weeks long. Average frequency can be computed by
dividing the total number of impressions by the total reach. An
average frequency of 2.5 means that, on average, the members of an
audience who have seen an advertisement have seen it 2.5 times over
a specific period of time. Media plans often specify a target
average frequency for a proposed advertising plan or schedule.
Adding more spots to an advertising plan or schedule naturally
increases the average frequency.
[0396] It should be recognized, however, that there is a tradeoff
between reach and frequency. Most known optimization methods suffer
from this tradeoff. For example, some combinations of advertising
campaigns tend to expose the same group of people over and over
again. This results in increasing frequency for those individual
who are exposed, but has only a limited effect on reach. Other sets
of advertising spots tend to expose many people a relatively few
number of times, increasing reach but not frequency. While adequate
in some circumstances, there are two limitations to using an
average frequency approach to specify the objectives of an
advertising campaign. The first stems from the fact that plans or
schedule with identical average frequency values can have very
different frequency distributions.
[0397] The data shown in the table in FIG. 20 illustrates a simple
example of two advertising plans or schedule that have identical
average frequency values, but with greatly differing frequency
distributions. Two members of the sample for plan or schedule A
were exposed only one time each, while person number 3 was exposed
seven times, thus resulting in an average frequency of three. The
audience exposure for plan or schedule B is ideal. Each member of
the sample is exposed three times, also resulting in an average
frequency of three. While these two plans or schedule represent the
two extremes of frequency distribution, these plans or schedule
demonstrate that there are media plans which have identical average
frequency values, but with significantly different frequency
distributions, and probably have differing influence on audience
members.
[0398] The other limitation of average frequency for optimization
purposes lies in the potential for having exposure timing patterns
at an individual level which are not optimal. Exposure to several
advertisements clustered during a short period of time followed by
a lengthy period with no exposure does not have the same value as
being exposed to a similar number of advertisements that are evenly
spaced over the entire period. The average frequency and the
frequency distribution for two different advertising plans or
schedule could be identical, but the average recall at the
individual viewer level could be dramatically different. The data
plot in FIG. 21 illustrates this point.
[0399] Assume that an advertisement is aired every day during a
four week period. The distribution of this advertising plan or
schedule is perfectly distributed. In this example, person A and
person B are both exposed to a total of four of the 28
advertisements aired during the period. Both are initially exposed
on day 1, which, for both, results in a recall rate of 30%. Person
A continues to be exposed for three more days in succession. The
incremental effect of these repeated exposures in recall ability
declines as indicated by concave shape of the curve. Each time
person A is exposed to another advertisement in quick succession,
the additional benefit in recall ability drops as the person
approaches a point of saturation. This is consistent with learning
theory, as well as with many studies for recall following exposure
to advertising. By contrast, person B is exposed once a week for
the next three weeks. Each time person B is exposed there is a
sharp rise in the rate of recall. In both cases, on the days when
the people are not exposed, their recall ability declines at a
steady and equal rate.
[0400] The recall results of the two exposure patterns are
significantly different. From day 1 to day 7, the recall ability of
person A is consistently above that of person B. But beginning with
day 8 until the end of the four week period the recall of person A
is consistently lower than person B. This simple graph demonstrates
that in order to maintain a steady level of recall, people need to
be exposed to advertisements consistently over time. Average
frequency, taken by itself, does not provide sufficient information
to determine how well specific advertising plans are meeting
objectives. Unfortunately, most presently known optimization
techniques rely heavily on average frequency to determine the most
optimal plan or schedule.
[0401] Effective Frequency
[0402] Effective reach refers to the total number of people who are
exposed to more than a specific number of advertising messages
(usually three) over a selected period of time. Summary frequency
distribution data is required to compute effective frequency. The
effective frequency tabulation for a simple advertising campaign
might be as shown in FIG. 22. In this case, out of 100 total people
in a small sample audience, there were 32 people who saw one and
only one of the advertisements, 15 who saw exactly 2 ads, etc. If
it is assumed that advertisements become effective only after
audience members have seen three advertisements, then the effective
frequency for this plan or schedule is 9. This is because a total
of 9 out of 100 people saw at least three advertisements.
[0403] The practice of using effective frequency to specify the
objectives of an advertising campaign suffers from the same
problems that average frequency does. While attempting to gauge the
value of a plan or schedule based on the total number of people
that might respond, plans or schedule with identical effective
frequency values can still have different frequency distributions,
and different advertising exposure distributions over time.
[0404] Weighted Effective Frequency
[0405] Weighted effective frequency valuation is an attempt to
account for the fact that all exposures may have some value, and
that plans or schedule which have skewed frequency distributions,
such as plan or schedule A shown in FIG. 20 above, are not as
desirable as are plans or schedule with more even
distributions.
[0406] To illustrate this point, different exposure values have
been assigned to different distributions as shown in FIG. 23. As
shown in FIG. 23, a value of 1.0 has been assigned to instances in
which an audience member is exposed to an advertising message for
the third time during a fixed period of time. It is important to
note that the second exposure to this same message has some value,
but may not as valuable as the third exposure. Therefore, the
second exposure in the fixed period of time is assigned a value
which is 80% of the value of the third exposure. The interpretation
of this value might be that if a person has an X% probability of
purchasing a product as a result of being exposed to an
advertisement for the third time, then this same person will have
an 0.8X% probability of purchasing the product as a result of the
second exposure. By the same reasoning, the first exposure has 50%
of the value of the third exposure.
[0407] The total value of multiple exposures can be calculated by
combining the value of each individual exposure. For example, if a
person is exposed to two advertisements which are both part of an
advertising campaign, the total value of both exposures is the sum
of the individual exposure values: 0.5+0.8=1.3. In theory, this
person has a 1.3X% probability of purchasing a product as a result
of being exposed two times to an advertisement. Now, if a second
person who has not seen any advertisements in the campaign joins
this first person, and the two of them are exposed to another
advertisement, then the value of this exposure is different for
each person. For the first person, the one who has already seen two
advertisements, the value of the third exposure is 1.0, while the
value of the same exposure to the second person is only 0.5. The
total value for these two exposures is 1.5. Using this method, the
total value of all exposures for plan or schedule A in FIG. 20 is
4.3 C two people are exposed only once for a value of 0.5 each,
while one person was exposed seven times for a total value of 3.3.
All exposures after the fifth exposure for person number 3 are
worth nothing. Using the same process, the total value of plan or
schedule B is 6.9. These total scores are more intuitively
appealing, and probably more accurately reflect the true value of a
given advertising plan or schedule.
[0408] Frequency Level Valuation Alternatives
[0409] The exposure level values in FIG. 23 are used only to
illustrate how to compute the total exposure value for an
advertisement. The assignment of value to various exposure levels
has been the subject of much debate over the years, and many
individuals have proposed exposure valuation schemes which are
significantly more complex than the one illustrated in FIG. 23.
These proposals are all rooted in an effort to understand how
audience members react to being exposed on multiple occasions to
advertisements. Five of the most widely discussed approaches for
valuing multiple exposures are reviewed below.
[0410] Linear
[0411] One simple approach to computing the value of exposure is to
assume that all exposures are of equal value; that the first
exposure to an advertisement by an individual viewer has the same
influence on the person as the twenty-third exposure. Under this
assumption, it is hypothesized that the more a person sees an
advertised product, the more likely he or she will be to buy it. If
this effect can be assumed to be perfectly linear; a person who has
seen an advertisement 101 times is 1% more likely to buy the
advertised product than a person who has only seen the
advertisement 100 times.
[0412] While this theory has obvious flaws, it is useful to present
as a contrast to other assumptions that are introduced below. FIG.
24 shows a plot of this "linear" assumption together with plots for
other assumptions. The horizontal axis is the exposure number,
i.e., the number of times that individual viewers have seen
particular advertisements. The vertical axis is the value of each
of these exposures.
[0413] As shown in FIG. 24, the value for each "linear" exposure is
equal to 1.0, regardless of how many previous times the viewer has
seen the advertisement. The twentieth time a viewer sees an
advertisement is just as valuable in persuading him to purchase the
advertised product as the first or second time he or she sees the
ad.
[0414] For purposes of optimizing an advertising plan or schedule,
no units of measure are required for exposure values. The
magnitudes indicate only the relative value of exposures. In the
linear case, they indicate that all exposures are of equal value,
but they do not indicate what that value is. Any number, such as
two or ten or any other number, could be selected as the constant
value. As will be shown below, the conclusions reached in
optimizing a plan or schedule will remain the same regardless of
the absolute value selected. While the relative values of
alternative changes in an advertising plan or schedule can be
compared, there is no way to compute the absolute value of an
advertising plan or schedule or a change in a plan or schedule.
[0415] Krugman
[0416] Many advertisers believe that a certain theoretical number
of advertising exposures is required to convert individuals into
purchasers. As an extension of this belief, media plans will
frequently specify required effective frequency values. These media
plans assume that if individual audience members are exposed to
fewer advertisements than this critical number during a given
period of time, then it is likely that the message will not be well
enough understood or sufficiently motivating to result in any
change in purchasing decisions. If, by contrast, audience members
are exposed to more advertisements than this critical number during
the same period of time, any exposures in excess of the critical
number will not have any additional influence because the audience
members will have already made their purchasing decisions.
[0417] This critical number of exposures has also been the subject
of much debate. Some scholars have suggested that the optimal
exposure number is universal and well defined, and that it is
independent of the product being advertised and the message
content. Others have argued that all exposures have some influence
on the purchasing decisions of audience members, but that the level
of influence follows a curve.
[0418] One of the most frequently cited studies is by Krugman.
Krugman claims that in order to influence people to make specific
purchasing decisions, they must be exposed to three and only three
advertisements. According to Krugman, with fewer than three
exposures, people will not yet be sufficiently aware or informed of
the product to consider making a change in their buying decisions.
Any exposure beyond three, Krugman claims, will have no influence
because people will have made their decisions, and that these
decisions are final.
[0419] This assumption about viewer behavior is illustrated in FIG.
24. The Krugman line is at zero for all exposure numbers except the
third. The total benefit of exposure number three is valued at ten.
Again, the actual value is arbitrary, but indicates that the value
of the third exposure is infinitely greater than that of all other
exposures. Beyond three exposures, exposure value for the Krugman
assumptions returns to zero.
[0420] S-Curve
[0421] It has also been suggested that the response to repeated
exposures of an advertising message is somewhat consistent with
Krugman's claims, but follows a gradual curve rather than an abrupt
spike as Krugman proposed. This theory postulates that people do
require several exposures to become fully aware and informed about
a product, but they also believe that the benefit of additional
exposure continues, possibly because of the need for periodic
reinforcement of the message. This enhanced benefit, however, comes
at a declining rate.
[0422] This curve is often referred to as the S-curve. The S-curve
is also illustrated in FIG. 24, which indicates the accumulated,
rather than the individual (as in FIG. 21), value of exposures for
an audience member. When the total number of exposures is low, the
S-Curve gradually increases in slope, indicating, for example, that
being exposed on four occasions to an advertising message has more
than twice the value of being exposed only twice. As the total
number of exposures becomes progressively greater, above about
seven in this case, the S-Curve gradually decreases in slope. The
value of being exposed 20 times is still greater than the value of
being exposed 19 times, but only marginally so.
[0423] The plotted value is referred to as an S-curve because, as
shown in FIG. 25, it does have a slight S shape. When plotted on an
exposure value curve as in FIG. 24, however, it is not S shaped. It
rises to a maximum point, which, in this case, is at about five
exposures, and then gradually returns zero as the number of
exposures increases.
[0424] Note that in each case, the curves in FIG. 24 are the first
derivative of the curves in FIG. 25. This means that the
accumulated area under each curve in FIG. 24 is the amplitude of
the corresponding curve in FIG. 25.
[0425] Diminishing Returns
[0426] As suggested by the S-curve, it is reasonable to assume that
being exposed to increasing numbers of advertisements follows a
line of diminishing returns. Some theorists have suggested,
however, that the first exposure is the most influential, and that
the value of subsequent exposures consistently drops. This is
represented in the diminishing returns line in FIG. 24. This line
is initially quite high, and then consistently falls with each
succeeding exposure. This indicates that the value of exposures can
become very small with large total numbers of exposures, but the
value never reaches zero. In terms of accumulated value of
exposure, FIG. 25 shows the curve increasing at a decreasing rate,
and approaches a limit at high exposure levels.
[0427] Eventual Irritation
[0428] Closely related to a belief in the law of diminishing
returns is the belief held by some scholars that people may
eventually become irritated with being exposed to the same
advertisement many times. As indicated in FIG. 24, this line at
some point falls below zero, indicating that the probably that
people will be converted into purchasers actually decreases at high
exposure levels.
[0429] Exposure Value Computation
[0430] For purposes of optimizing an advertising campaign, the
value of exposure is most conveniently represented by individual
exposure value as shown in FIG. 24 rather than by the total value
of accumulated exposures as shown in FIG. 25. The exposure values
for FIG. 24 are summarized in the table shown in FIG. 26.
[0431] In planning an actual advertising plan or schedule, exposure
values may not be known in detail, nor may there be a high level of
confidence in the values that are known. However, any information
that the media planner may have, either from technical journals,
research, corporate experience, or just gut feel, concerning the
value of additional exposure is useful in optimizing an advertising
plan or schedule. The optimization methods of the present invention
provide a mechanism for incorporating whatever information is
available into the decision making process in a systematic way.
[0432] Referring to equation (1) above, the total score due to the
value of exposure for an audience sample is: 19 i = 1 N V I n ( i
)
[0433] where
[0434] N Total number of exposures of advertisement a for an
audience
[0435] V.sub.I.sup.n(i) The value of exposure i which is the
n.sup.th exposure of advertisement a of a member of the
audience.
[0436] To compute the total value in equation (2) for the addition
of one spot to an advertising plan or schedule, the following steps
are performed. (See the table in FIG. 27 for a sample of the
computations for an audience sample of 15 people.)
[0437] 1) Tabulate the frequency for each member of the sample for
the advertising plan or schedule without the additional spot. (The
second column in FIG. 27)
[0438] 2) Identify those people who saw the additional
advertisement. (The third column in FIG. 27)
[0439] 3) Enter the individual value per exposure from FIG. 24 or
FIG. 26 for those who saw the additional advertisement based on the
exposure valuation assumptions.
[0440] 4) Total the individual exposure values.
[0441] The relative value of the total scores between, Krugman and
the S-Curve is not important. As previously mentioned, the value of
10 for the one exposure for person #6 under the Krugman assumption
is an arbitrary number. Any value could have been selected. The
objective is not to compute the total value of an advertising plan
or schedule, but to compute the change in value of a plan or
schedule for alternative plan or schedule modifications. For
example, using the S-Curve, a score can be computed for a
particular advertising plan or schedule plus one added advertising
spot. Then using the S-Curve again, the score can be re-computed
for the same advertising plan or schedule, but with a different
spot added to the plan or schedule. These two scores can then be
compared to determine which additional spot contributes the
greatest value to the plan or schedule based on the S-Curve
exposure values. It is possible to reach an entirely different
conclusion on how to optimize the plan or schedule if other
exposure valuation assumptions were adopted, such as the
assumptions inherent in the Krugman curve.
[0442] Using Exposure Values to Optimize Advertising Plans or
Schedules
[0443] While the simple examples shown above might work for an
audience of 10 people, there are additional considerations once the
size of the audience expands to larger levels.
[0444] The graph shown in FIG. 28 illustrates how exposure
valuation can be used to gauge the relative value of two
alternative advertising spots. For example, assume that for a
particular advertising campaign, exposure levels between two and
six exposures are ideal and of approximately equal value, while
exposure levels less than two or greater than six are worth
nothing. In addition, assume an existing advertising plan or
schedule to which one additional spot must be added. The frequency
curve for the base plan or schedule is shown in FIG. 28. On top of
this frequency curve is plotted the change in frequency that would
result from adding either of the two alternatives to the base
plan.
[0445] As indicated in FIG. 28, the change in frequency for both
spots ranges from one to about 15. In both cases there are people
who are exposed few times, and people who are exposed many times.
However, the average change in frequency for spot B is clearly less
than change in frequency for spot A. The plot for spot B is thicker
at low numbers of exposures than is the plot for spot A. In the
range of from two to six exposures, the sum total value of exposure
for spot B probably exceeds the total for spot A, even though the
total number of exposures for spot A may be slightly greater than
the number of exposures for spot B.
[0446] Even though spot B appears preferable to spot A, note that
spot B is not ideal in the sense that there are many exposures
which fall outside the range considered to be valuable. An ideal
advertising campaign would be one in which all audience members are
exposed to precisely the specified number of exposures. But no
campaign is ideal. If, however, one spot were to be added to our
base campaign, and if spots A and B are the only two spots
available, then adding spot B would be the optimum solution, even
though it is not an ideal solution. The process of optimizing an
advertising campaign is one of selecting from among the available
spots those spots that maximize the total value of all exposures
for the campaign.
OTHER ISSUES
[0447] Reach
[0448] One important criteria against which many advertising
campaigns are measured is the total number of people who are
exposed to one or more advertising messages over a specific period
of time. This is termed reach. As discussed above, is has been
observed that modifications to an advertising plan or schedule to
increase either reach or frequency is often at the expense of the
other. In addition, because of the limitations associated with
reach, the underlying wisdom of using reach as a measure for the
value of an advertising campaign has been questioned.
[0449] If a media planner were to optimize a plan or schedule based
on total reach using the weighted effective frequency method, then
the Krugman curve with the spike at one exposure (see FIG. 24)
would accurately describes this objective. Using this weighting,
only when audience members are exposed to advertisements for the
first time are the exposures included in the total exposure
valuation. This is precisely the definition of reach. In other
words, for a weighted effective frequency exposure valuation table
similar to the table shown in FIG. 23 for optimizing a plan or
schedule based on reach alone, the values would be as shown in FIG.
29.
[0450] Effective Frequency
[0451] If a Krugman curve with the spike set at three exposures is
adopted, the results for the weighted effective frequency valuation
method become identical to the commonly known effective frequency
valuation method. Exposures below three are worth nothing, as are
all additional exposures above three. The weighted effective
frequency optimization method presented herein is a more general
type of effective frequency. Using a simple Krugman curve, the
generally accepted version of effective frequency can be derived.
Using one of the other frequency valuation curves, such as the
S-Curve, a continuous frequency function can be defined, which, in
effect, describes the probability of decision and conversion at
multiple frequency levels, a somewhat arguably more realistic
assumption. This approach recognizes that there is no clearly
defined point of conversion as Krugman claims, but allows the
valuing of exposure at many levels.
[0452] There are many other scoring alternatives that are more
accurate in describing the value of advertising exposure than
either reach or conventional effective frequency. However, the
optimization methods of the present invention as described herein
are flexible enough to allow optimization of an advertising plan or
schedule based on any set of criteria for exposure value
measurement, including reach and simple effective frequency. It
should be noted, however, that the most preferred embodiments use
additional valuation techniques to further improve and refine the
results.
[0453] Promotions
[0454] A television program promotion is an advertisement on a host
program which publicizes a target program. The intent of a given
promotion is to increase the probability that audience members will
choose to watch the targeted program. In other words, the objective
of promotion could be either of the following: to persuade those
people who otherwise would not have watched the target program to
watch it; or, to encourage people who generally do watch the target
program not to defect to other programs. Promotions that target
only those who are not currently loyal viewers could be distinctly
different than promotions that target loyal viewers. They could be
made more introductory in nature, and might possibly be somewhat
longer, while loyal viewers might only require brief teasers to
maintain their interest and commitment to watch.
[0455] It can be shown that for regularly scheduled weekly
programs, relatively few people alternate between competing
programs from one week to the next. They are either loyal or not.
Very few are undecided one week after another and alternate between
competing programs. Therefore, it may possible to promote programs
using promotions which are targeted to any one of three groups:
groups of loyal viewers; groups of non-loyal viewers; and groups
which have both loyal and non-loyal viewers. If viewers are
partitioned for a program according to their level of program
loyalty and promotions are created accordingly, then the
optimization is slightly different for each of the
alternatives.
[0456] Referring back to FIG. 27, for conventional advertisements,
the exposure scores for members of the sample audience are included
only if they viewed the advertisement. For promotions aimed at
loyal viewers, the process is modified slightly by including
exposure scores only if the audience members satisfy two
conditions. The audience member must both see the promotion and see
the target program. For a promotional campaign aimed only at
non-loyal viewers, to be included in the scoring totals, the
audience members must both see the promotion and not see the target
program.
[0457] Finally, the optimization process for promotional campaigns
which are aimed at both loyal and non-loyal viewers can treated as
conventional advertisements in which audience viewership (i.e.,
whether or not the audience members saw the target program) is
ignored. What makes it possible to modify the process for
promotions in this way is the fact that the data that are used to
measure exposure are the same data that are used to measure the
response. This is the data contained in DME database 126. For
purposes of promotion analysis, person-by-person exposure data
provides a form of single source data. If single source data was
available for other advertised products, then this type of analysis
would also be possible. Under some circumstances or for some types
of products, one could choose to exclude exposure values for all
audience members except for potential customers. People who are
already loyal users of the advertised product would not contribute
to the optimization process.
[0458] In the optimization process, if the targeted audience is
determined based on whether or not they are loyal viewers (or loyal
customers), then second-order effects, which are not otherwise
present, can be introduced into the optimization process. This
means that the optimal plan or schedule could change over time as a
direct result of executing the optimal plan or schedule itself.
This occurs through the following sequence of events: 1) create a
promotional campaign which is aimed at, for example, loyal viewers
only; 2) select a given set of media objectives and assumptions
about the role and effectiveness of promotions; 3) using the media
objectives, arrive at a particular optimal plan or schedule; 4) use
the optimal plan or schedule for airing the promotions; 5) the
nature of the promotion plan or schedule influences a change in the
viewing patterns of some members of the audience--these members,
over time, either remain loyal to the target program because of
exposure to the promotions, or defect to competing programs because
of the lack of adequate promotion; 6) the changes in viewing
patterns then change the results of the optimization process for
future promotions. This occurs because individual exposure values
are included in the audience exposure totals based on whether or
not the are loyal to the program. If the people who see the target
program change, then the total scores and the resulting optimal
plan or schedule will also change. Similar logic applies to a
second-order effects in conventional advertising campaigns if an
advertising campaign can be targeted exclusively at loyal or
non-loyal customers. Note that the methods of the present invention
as shown in FIG. 13 provide a feedback mechanism to the overall
process to allow a media planner to account for and adjust for
second-order effects.
[0459] Additional research may be performed to identify valid media
objectives for promoting programs. This research will typically
require designed experimentation with promotions, and analysis of
the resulting viewing behavior. It should be noted that the methods
of the preferred embodiments of the present invention described
herein are sufficiently broad to accommodate any type of media
objective that may be derived from a study involving promotions. In
addition, a computer system using the preferred optimization
methods and techniques of the present invention as described herein
is a valuable tool in aiding in the analysis of this type of
experimentation. Even without any experimentation, the preferred
embodiments of the present invention are useful in improving the
efficiency of existing promotion plans or schedule based on the
planning assumptions and spending levels that are already in
place.
[0460] Time Weighted Effective Frequency
[0461] Time weighted effective frequency valuation is another new
valuation technique that goes one step beyond the techniques used
for weighted effective frequency. Here, in addition to recognizing
the fact that all exposures may have value, as weighted effective
frequency does, it is also recognized that the distribution of
advertising exposure over time is important in gauging the reaction
that audience members will have to an advertising plan or schedule.
As shown in FIG. 21, a cluster of exposures may not be as desirable
as a group of more evenly distributed exposures would be.
[0462] Ideal advertising exposure, then, involves: 1) providing
individuals with an adequate number of exposures; 2) over a
specified period of time, 3) with the proper spacing between
exposures. Optimizing an advertising plan or schedule to achieve
this objective is not possible with any known techniques. But with
individual exposure data extracted from database 126, it is
possible to identify plans or schedule from among various
alternatives that have optimal exposure patterns for individuals,
not just optimal exposure frequencies.
[0463] Before examining some of the techniques that may be used to
compute the time weighted effective frequency for an advertising
plan or schedule according to the preferred embodiments of the
present invention, it is necessary to explore in more detail the
influence of timing in advertising. Specifically how people learn
and forget over time.
[0464] Learning and Recall in Advertising
[0465] The first time people are exposed to advertising, their
ability to recall is less than perfect. The second time they are
exposed, their recall improves. The incremental amount that they
learn as a result of the second exposure, however, may not be as
much as the amount learned from the first exposure. With a third
exposure, recall shows additional incremental improvement, but to a
lesser extent than exhibited with the first exposure.
[0466] This phenomenon is illustrated in FIG. 30. Three groups of
people are exposed to advertisements for three separate products on
seven successive days, followed by 21 days in which they experience
no advertising exposure. The curves shown in FIG. 30 plot the
average ability for audience members to recall the advertisements
for each of the three products. Each of the three curves rises
steeply for the first two or three days, but then each begins to
level out. Each successive exposure results in a recall
improvement, but the improvement diminishes with the number of
exposures. In each case, the recall rate approaches a theoretical
limit, which in this case is 60% for each of the three products.
This simply means is that no amount of advertising will raise the
recall rate above that level.
[0467] The difference between the three curves for the three
products is a reflection of a variety of factors that might
influence the rate of learning, such as advertisement length,
product value, etc. These issues will not be examined here.
[0468] Note that although the learning curves for the first seven
days in FIG. 30 are similar in shape to the curve for diminishing
returns in FIG. 25, the audience behavior assumptions behind the
two curves are not the same, and should not be confused.
Diminishing returns as described in FIG. 25 refers only to
frequency rates as indicated by the "number of exposures" scale on
the horizontal axis. The learning curve in FIG. 30, by contrast,
introduces a factor of time into the influence of advertising. It
suggests that diminishing returns occurs only when exposures are
closely spaced in time. Refer to FIG. 21 for an example where two
exposure plans or schedule with identical frequency levels result
in significantly different influence curves over time. Curve A
clearly shows the effects of diminishing returns while curve B does
not.
[0469] Recall Decay
[0470] If advertising is not reinforced with additional exposure, a
person's ability to recall an advertised product declines over
time, as does influence that promotions have on television viewing
decisions. A number of studies have found that recall and the
decline in recall over time is dependent on a variety of
environmental and demographic factors such as product category,
time of day, host program, personality, and exposure history.
Referring again to FIG. 30, beginning on day 8 the advertising
message is no longer reinforced with additional daily advertising
exposures. When reinforcement ends, the recall curves immediately
begin to decline. The rate of decline in average recall for the
three products differs, but in each case the curves consistently
drop towards zero. Initially the slope for each of the curves is
steep, but becomes less so over time.
[0471] Influence Index
[0472] For purposes of scoring advertising plan or schedule changes
using time weighted effective frequency, the vertical axis in FIG.
30 is redefined to be the influence index, which is the relative
level of influence that exposure to one or more advertisements in
an advertising plan or schedule has on purchasing decisions. The
minimum value is zero, which means that previous exposures to
advertisements in an advertising plan or schedule have no influence
on current purchases. The maximum influence index value is 100%. If
a person is at a 100% level of influence, it means that the
influence on purchasing decisions which results from being exposed
to advertisements will not increase with additional exposure to the
advertising.
[0473] Influence index as defined herein should not be interpreted
as the probability of purchase. It is only an index that indicates
a level of influence if a person makes a purchase. People at 100%
influence index do not necessarily buy the advertised product when
given an opportunity. It only indicates that the level of influence
cannot increase with additional exposure.
[0474] Adopting this convention allows us to describe the curves in
FIG. 30 logarithmically using only two parameters:
[0475] .alpha. The rate of increase in influence with exposure
[0476] .beta. The rate of decline of influence without exposure
[0477] The equation for the influence index at time t+1 is: 20 I t
+ 1 = { I t + ( 1 - I t ) Was exposed ( 1 - ) I t Was not exposed (
3 )
[0478] and the plot where .alpha.=0.4 (suggesting a moderately
rapid improvement in influence), and .beta.=0.1 (a gradual decline
in influence) is shown in FIG. 31. This person is exposed to
advertisements on seven consecutive days. Then, beginning on day 8,
the person is exposed to no other advertisements through the end of
a four week period.
[0479] This plot is generated by computing the influence index for
each day in succession using equation (2) above. Beginning with day
1, a day in which the person was exposed, the influence index is
computed as:
I.sub.1=0.0+.alpha.(1-0.0)=0.0+(0.4)(1.0)=0.40
[0480] This same person on day 2 is exposed to another
advertisement. The resulting influence index on day 2 is:
I.sub.2=0.40+.alpha.(1-0.40)=0.40+(0.4)(0.60)=0.64
[0481] If the person at day 7 is at an influence index of 97%, and
is not exposed to an advertisement on day 8, then the resulting
influence index on day 8 is:
I.sub.8=(1-0.10)0.97=0.87
[0482] The advertising frequency required to maintain a given level
of influence depends only on .alpha. and .beta.. If, under a given
set of market and product conditions, the level of influence grows
slowly and/or declines quickly for an advertising message, then the
message requires frequent reinforcement. If the influence increases
quickly and/or declines slowly, then reinforcement is required less
frequently. It should be noted that the alpha/beta approach to
modeling learning and decay is only one possible alternative for
modeling the concepts of learning and decay. Those skilled in the
art will recognize that many other techniques may be applied to
accomplish the same or similar results. For example, a numerical
look-up table with a series of predetermined values that have been
empirically derived from a series of marketing studies could also
be used to model learning and decay. The scope of the present
invention contemplates the use of alternative modeling
techniques/methods and includes all such similar techniques.
[0483] Advertising Influence Over Time
[0484] The total influence that a series of advertisements has on
an individual accumulates over time. When there is either a
sequence of exposures to an advertisement, or periods of
non-exposure, as shown in FIG. 31, the curves are smooth and
consistent. When exposure is intermittent, the curves are also
logarithmic, but in a discontinuous way. This is illustrated in
FIG. 32. In this case two individuals are initially exposed
unequally to an advertising plan or schedule. Beginning with day
number 1, Person A is exposed to three consecutive advertisements.
Then, on day 4 and beyond, both person A and person B are exposed
to the same advertisements B on day 4 and again on day 8.
[0485] As indicated by FIG. 32, even two weeks after the sequence
of advertisements began, there continues to be a difference in
total influence for the two people. Person A is at 19%, while
person B is at 16%. The difference between the two people is small,
and gradually becomes smaller over time, but it continues to be
measurable. This phenomenon is entirely consistent with most
theoretical learning models. Even though the exposure history for
both people has been identical for many days, there continues to be
some residual influence that stems from earlier exposures for
person A. As time progresses, the difference consistently
decreases, but it never entirely disappears.
[0486] Illustration of Plan or Schedule Change
[0487] Modifying an advertising plan or schedule to more evenly
distribute advertisements may not improve the overall effectiveness
or value of the plan or schedule if the individual exposures do not
become more evenly distributed as a result of the modification. It
is possible that a plan or schedule could be evenly distributed
over time, but still have an overall uneven distribution of
exposure because of uneven distribution of exposure by various
categories of audience members.
[0488] To illustrate how a plan or schedule might be modified to
achieve a better time distribution of exposure without necessarily
modifying the distribution of the actual advertisements, a more
realistic example, at least in terms of how people might actually
be exposed to advertisements in a plan or schedule, is presented
below.
[0489] Referring now to FIG. 33, three groups of audience members
are each represented by an average influence index. The three
groups may constitute only a small percentage of the total
audience, they may be of different sizes, and they may not follow
clear demographic, or geographic boundaries, but the exposure
decisions within each group are moderately consistent. During the
four-week period shown, group A is exposed only twice; once at the
beginning of the period, and again at the end of the period. For
most of the period, the influence score for this group is
relatively low. Group B is exposed repeatedly during the entire
period, and, as indicated by concave curve in two places, has
reached saturation. The exposure for Group C is more ideal in that
the overall exposure is relatively consistent, and the resulting
influence values are about 40%.
[0490] The strategy used to optimize the plan or schedule for these
three groups of people is to adjust the general advertising plan or
schedule to improve on individual timing irregularities. If the
plan or schedule can be rearranged so that one or more of the
excess exposures for group B can be shifted to group A, then the
total influence index value for the audience as a whole increases.
So, in an attempt to optimize this plan or schedule based on
exposure timing alone (ignoring audience valuation, cost, response,
etc.) one of the advertising spots could be changed using the
following steps:
[0491] 1) Go through the entire advertising plan or schedule and
find the one spot which contributes the least to the total
influence index value for all audience members combined. Remove
this spot from the advertising plan or schedule.
[0492] 2) Assemble a list of alternative spots that would be
acceptable additions to the plan or schedule. This list could
conceivably include all available spots during the four week
period, but could also be confined to a relatively few alternatives
in competing vehicles.
[0493] 3) Go through the list of spots and find the one which
contributes the most to the total influence index value for all
audience members combined. Add this spot to the plan or
schedule.
[0494] 4) Repeatedly go through steps (1) through (3) until no
further improvements can be made to the plan or schedule.
[0495] In the above procedure, the three viewing groups are not
explicitly identified. The purpose of the previous discussion was
to simply highlight the potential for improving an advertising plan
or schedule based on the exposure habits of the three groups. It
should be noted, however, that any change in the schedule which
results in shifting of exposures from group B to group A without
disturbing the exposures for group C will result in an improvement
in the total influence index. All possible changes should be
examined to identify those which will result in an improved
index.
[0496] Scoring
[0497] In step (3) above, the total influence index value for all
audience members combined is computed. Based on time weighted
effective frequency, the value of an advertising campaign is the
total influence that the advertisements have on individual audience
members. The influence that the campaign has on an individual is
the sum of the influence for each day during the campaign. This is
explained in more detail below.
[0498] FIG. 32 duplicates the curves in showing the influence on
individuals for two different advertising exposure sequences. Both
have a frequency level of four, but the exposures for person A all
occur during the first four days of the period, while the exposures
for person B occur evenly throughout the four week period. The
vertical axis for this figure has been labeled "influence index"
and is scaled from 0% to 100%.
[0499] Looking at FIG. 34, an effective way to measure the
influence of advertising on an individual over a period of time
would be to measure the average index over the period (or,
equivalently, to measure the area under the curve.) The average
index for person A is 679/28=24%, while for person B it is
875/28=31%.
[0500] With individual exposure data, the only other information
which is required to compute time weighted effective frequency is a
specification of the shape of the learning and decay curve using
two parameters: .alpha. and .beta.. The methods of the present
invention for optimizing an advertising plan or schedule using time
weighted effective frequency do not necessarily require that the
level of influence of advertisements be described
logarithmatically. However, this is a simply a mathematically
convenient technique for characterizing learning and recall. If,
for instance, it was determined that the audience response for a
particular advertising campaign more closely followed some other
function, such as a step function, then the time weighted effective
frequency technique of the present invention could accommodate that
belief. The methods used would remain the same.
EXAMPLE
[0501] FIG. 35 illustrates how scores are computed for time
weighted effective frequency using actual exposure data. Shown is a
base plan or schedule spanning one week beginning on day 1 and
ending on day 7. The plan or schedule for each of the seven days,
with the exception of day 4, has a single advertisement. To expand
the plan or schedule and add one more spot to the on day 4, various
alternatives may be considered. For purposes of illustration,
consider one of three alternatives: 8:00 PM, 9:00 PM, or 10:00 PM.
The audience is composed of a total of 20 people. The exposure
columns indicate whether or not each of the people in the audience
was exposed to the advertisement on each of the indicated days,
including the three alternative time spots on day 4.
[0502] In examining the exposures, there are no obvious patterns in
the viewing. Some audience members view more television than
others, some are not exposed at all during the week, and others
have only one or two exposures. The three alternative spots are
approximately the same in the sense that all three have a total of
five audience members who would be exposed.
[0503] The influence index values for each of the 20 people are
computed for each of the three alternatives. The values chosen for
this example for .alpha. and .beta. for all three alternatives are
0.4 and 0.1 respectively. If, by a particular day, the person has
not yet been exposed, then the index is zero and is not listed. If
a person is exposed on any of the days, then the index for that day
increases. If a person is not exposed, the index decreases. Person
#14, for example, is exposed to the advertisements on day 4 for
both alternative A and alternative B, but not for alternative C.
The influence index values for alternatives A and B are the same
and both increase on that day. After day 3 the alternative C scores
are less than either alternatives A and B because person #14 was
not exposed to an advertisement at 10:00 PM on day 4. These scores
for person #14 are plotted in FIG. 36.
[0504] Returning to FIG. 35, if a person is not exposed to any of
the three alternative spots, such as person #20, then that person's
computed scores are identical for each of the three alternatives.
For all members of the audience, the scores for the first three
days for each of the three alternatives are identical.
[0505] The influence index values are totaled by person, and then
for each of the three alternatives, a grand total is computed which
is the sum of all influence index values for all audience members
over the 7 day period. If the only consideration is exposure
influence (without demographics, cost, etc.) then alternative B
appears to be the superior alternative. At 3729, alternative B is
about 1.5% better than alternative A and 1.2% better than
alternative C.
[0506] A difference of 1.5% may not seem significant, but may be
very significant given the number of constraints in this example
which limits the amount of optimization that can be done,
including: considered only scheduling in the optimization (i.e.,
ignoring demographics, cost, response, etc.); only seven
advertising spots; spanning only 7 days; only one change; only
three alternatives spots; approximately equal exposure to spots;
and only 20 members in the audience sample.
[0507] Even more significant improvements in advertising efficiency
can be found as it becomes necessary to address more realistic
problems where there is additional flexibility in iterating many
times through the optimization procedure with more possible changes
in the plan or schedule and more alternatives for each of the
changes.
[0508] Even with this simple example, computing the total score for
each of the three options is a significant task. Computing the
scores for several dozen alternatives for a base plan or schedule
of one or two hundred spots and an audience of many thousands of
people obviously requires a computer system, such as the one
described herein. It is a computationally intense method, but the
only parameters that need to be specified are .alpha. and .beta..
An attractive feature of the methods of the present invention is in
the simplicity in arriving at worthwhile results.
[0509] Why Plans or Schedules are Often Not Optimal
[0510] Even if advertisements are evenly distributed over time, it
is possible that the resulting plans or schedule may still be less
than optimal. It is possible that individual advertising exposure
is not evenly distributed, even with an evenly distributed plan or
schedule. This phenomenon is described in more detail below, using
the information presented in the optimized plan or schedule sample
above.
[0511] Referring again to FIG. 35, in the audience of 20 people,
there were three people, #1, #9, and #17, who were not exposed on
either day 2 or day 3, but who were available for exposure on day 4
at 9:00 PM. Two of these three people were exposed on day 1, but
the other was not. By day 4, all three of them had relatively low
index values. Their influence index totals would contribute
substantially to the grand total if they all could be exposed on
day 4. In relative terms, this group of three people was a
significant collection in the audience sample. They had similar
viewing habits, and they all needed another exposure to maintain
awareness and to increase the value of the advertising message to
the advertiser.
[0512] These three people were not necessarily the most lightly
exposed in the audience. Most others in the sample were not exposed
at all, or were exposed only once before day 4. The difference was
that this small group shared similar viewing habits. Although
unrelated, these disparate individuals made decisions about viewing
that were similar, and this similarity could be used to improve
total influence for the audience.
[0513] Therefore, even though the spots in an advertising plan or
schedule may be evenly distributed, the audience is not. The focus,
then, of the methods of the present invention is not on the overall
timing of advertising spots, but on fine tuning the plan or
schedule so that individual exposures are also optimally
distributed over time. It is possible to have an evenly distributed
advertising plan or schedule, and still have uneven individual
exposure to that plan or schedule simply by virtue of the fact that
certain groups of audiences members typically make similar media
selection decisions. This fact lends credence to the value of
optimization to increase not only the reach of an advertisement but
to more evenly distribute the audience exposure according to these
media selection patterns.
[0514] Given that some form of audience grouping is required for
optimizing plans or schedule, then identifying these groups and the
patterns associated within the groups within a larger audience
sample is important for advertising optimization at a national
level. Not only do identifiable groups of viewing patterns exist in
most audiences, but these group movements can dominate audience
exposure. In reviewing the viewing habits of the Nielsen Television
Index sample, for example, it is easy to find systematic decision
making patterns everywhere. The existence of these patterns stems
from several conditions.
[0515] First of all, certain types of programs appeal to specific
groups. These groupings may stem from a variety of complex,
interacting factors including gender, age, culture, geography,
weather, income, race, etc. Next, most people have strong program
loyalties. Although these program loyalty patterns can vary from
one program to another, and from one demographic group to another,
all programs command some degree of loyalty from their audiences.
In addition, lead-in and lead-out programming is a strong
determinant in program selection. Further, competing programs with
their audiences of loyal viewers and the surrounding programming
associated with these programs can influence viewing patterns
significantly. Finally, there are a host of less quantifiable
factors which can influence viewing patterns.
[0516] A combination of these factors often results in groups of
unrelated people consistently making similar media selection
choices over time. This condition can be used to the advantage of
an advertiser in one of two ways, depending on the media objectives
of the advertiser. For example, if the objective of the advertiser
is to simply maximize total reach, then an advertising plan or
schedule should be designed that will effectively expose each
identified group one time each, focusing the available resources on
broad coverage. However, if the objective of the advertiser is to
optimize frequency, possibly at the expense of reach, then large
groups should be exposed the specified number of times, according
to the model of optimal effective frequency as determined by the
media planner.
[0517] The process of optimizing an advertising plan or schedule
naturally takes advantage of decision groupings. If a certain group
requires additional exposure during some period of time, then the
optimization process identifies advertising spots which the target
group is frequently exposed to. If a group is over-exposed, then
the process also identifies spots which the group is collectively
exposed to and eliminates them from the schedule during the
optimization process.
[0518] The validity of using historical viewing records to measure
the potential for future advertising campaigns should be examined
at this point. It is the consistent patterns in individual and
group exposure that not only makes optimization possible, but also
justifies its use. Individuals and groups consistently follow
patterns of viewing. If a given viewing or other media-related
audience sample, such as the Nielsen sample, follows these types of
patterns, then the larger population also does. There are
indications, for example, that program loyalty patterns are very
stable. Once an individual has become loyal to a particular
program, this loyalty will often remain in place for months or even
years. If these patterns are in place, then the patterns can and
should be used to improve on advertising plans or schedule.
[0519] Previous Studies
[0520] Many researchers have attempted to use a variety of
techniques to identify patterns in television program selection.
These efforts have met with only mixed success. The difference
between previous attempts and the present invention lies not only
in the methods and techniques, but also in the objectives. Most
previous efforts were attempting to identify program types and
audience groupings so that new programming could be identified
which would appeal to these audience groups, so that rating and
share forecasts could be made, and so that programming niches could
be found for new television stations or cable channels.
[0521] In all of the studies, many audience groups were identified.
There were, however, several problems with the results of the
studies. For example, some studies have shown that there are too
many people in the television audience who do not fit neatly into
any group, thus making accurate forecasts difficult to compute.
Further, too many groups seem to be driven by schedule-dependent
factors, such as lead-in and lead-out, which lend little insight
into programming decisions. In addition, many of the groupings are
illogical. Program type preferences, for example, do not dominate
decision patterns. This may be explained by suggesting that people
actively seek out variety in programming. Alternatively, it is
simply possible that many decisions are based on the least
objectionable alternative available.
[0522] The intent of the methods and techniques of the present
invention is only to find ways to improve advertising plans or
schedule. The viewing habits of any group of people in the audience
can be used to make individual exposure spacing improvements. There
is no requirement for widespread, consistent grouping to
successfully implement the optimization techniques disclosed
herein.
[0523] Using the methods herein, groups do not need to be
explicitly identified. The techniques do not require modeling
(i.e., factor analysis, linear programming, or regression analysis)
of either the audience or the media as other methods have employed.
The techniques simply take advantage of audience decision groupings
to optimize an advertising plan or schedule by improving the
overall influence index score. The reason that improvements can be
made to advertising plans or schedules using time weighted
effective frequency techniques is because audience groups exist.
But the techniques do not require the isolation and identification
of these groups. Instead, the present invention uses a specific
methodology to search through a set of modifications exhaustively
at each step to find the optimal combination which most closely
match the desired criteria.
[0524] Ideal vs. Optimal Plan or Schedule
[0525] Returning to FIG. 35, the selection of alternative B is not
ideal, but it is the optimum alternative given the set of three
alternatives. Person number 14, for example, didn't require another
reinforcement of the advertisement as much as some of the other
people in the audience. The index for this person only increased
from 78 to 87; a total of 9 index points. This is well below
average for the people who saw the advertisement on day 4 at 9:00
PM. Alternative B provided the opportunity to expose person #9 who
had never seen the advertisement before, but missed person 19 who
also had never seen it.
[0526] Some audience members will continue to be exposed
irregularly, in part because of irregular viewing habits. Others
will be over- or under-exposed because of above or below average
viewing. While it is impossible to achieve an ideal distribution of
exposures for all viewers in all circumstances, the intent is to
optimize the distribution of exposures so that the exposure
distribution, on average for the entire population, is
improved.
[0527] Audience Valuation
[0528] As defined herein, the value of an individual to an
advertiser depends entirely on demographic characteristics. The
value of an individual does not depend on media habits, the number
or timing of advertising exposures, or the cost of advertising. The
value is based purely on the belief that specific demographic
characteristics are good indicators of how people will respond to
specific advertising. As mentioned previously, some demographic
groups generally buy some types of products. Some groups rarely
purchase consumer products, but when they do, their purchases tend
to be large.
[0529] Some demographic groups may rarely select a particular type
of entertainment. For example, studies have shown that rich
middle-aged men are difficult to reach with network television, but
their underlying value to advertisers is independent of this fact.
The question in valuing an audience member is, if it is possible to
reach a given person with advertising, how valuable would the
person be to an advertiser? The index values for each specific
demographic parameter or characteristic can be assigned based on
the importance of the parameters for a given advertiser.
Alternatively, previously identified values can be implemented in
the present system and used directly.
[0530] The value of an individual audience member to a particular
advertiser is defined in equation (1) above as 21 d = 1 D V a d = V
A 1 .times. V A 2 .times. .times. V A D
[0531] where, for example, V.sub.A.sup.1 might be the age/gender
index value, V.sub.A.sup.2 could be the household size index value,
etc. Each of these index values are multiplied together to
calculate the index value for an individual audience member.
[0532] Various statistical techniques may be used to draw clear
demographic boundaries between product purchasers and
non-purchasers. However, most known valuation techniques ignore the
fact that all members of an audience may have some value, but that
the valuation function must be a continuous rather than a discrete
one. While exposing a targeted audience to an advertisement, other
audience members who are not part of this targeted audience are
still exposed, and these exposures have some value.
[0533] The formatting used for the demographic data is similar to
the formatting used in valuating frequency levels. There can be any
number of partitions for each of the demographic characteristics
depending on needs and the information available to the advertiser.
The tables shown in FIGS. 37, 38, and 39 illustrate how various
sample values might be formatted for indexing purposes. As with
other illustrations, these tables are merely possible approaches
and the present invention is not limited to these specific
examples.
[0534] Using these assignments, a 24 year old women, living in a
home with $35 K annual income, in a B size county would be valued
at: 22 V A 1 .times. V A 2 .times. V A 3 = 0.6 .times. 0.7 .times.
0.8 = 0.336
[0535] Although not essential for the optimization method to
function properly, in this case the maximum index value for each of
the three demographic characteristics is 1.0. Therefore, a woman of
age 28-40 with a household income of 40+K living in an A county
would score: 23 V A 1 .times. V A 2 .times. V A 3 = 1.0 .times. 1.0
.times. 1.0 = 1.0
[0536] Therefore, the 24 year old woman has 33.6% of the value of
the most valuable person to this advertiser. To further the
example, assume that there are 10 people in a small audience
sample, and that five of these people saw the advertisement. Assume
that all five people are demographically identical to the 24 year
old woman described above. To compute the total value of this
audience to an advertiser, the scores of each of the 5 individuals
who saw the advertisement are summed: 24 i = 1 N a [ V I n ( i )
.times. d = 1 D V A d ( i ) ]
[0537] where N is the total number of impressions for this
advertisement, which in this case, is five. Audience members who do
not see an advertisement do not contribute to the overall total
score. So the sum of all five audience members is: 25 i = 1 5 .336
= 1.68
[0538] The range of demographic factors which can be used to define
audience member values can extend as broadly as there is data
available. Network television advertising contracts specify only
age and gender, but the analysis that precedes that contract
agreement can include any number of demographic factors, and could
even include value assignments to age and gender categories that
are not a part of the contract provisions. For example, an
advertising plan or schedule for a product with a narrowly defined
target group such as women 15-25, could be optimized based on that
group. Then, the advertising agreement could specify a more
conventional demographic group such as women 18-34.
[0539] As indicated in FIG. 37, the methods of the present
invention allow the assignment of values to combinations of
demographic characteristics. Women 28-40 are valued at 1.0, but men
in the same age range are worth only half that much to the
advertiser. Of course, as the number of demographic characteristics
increases, the number of possible combinations of demographic
characteristics increases exponentially.
[0540] There are a number of places that advertisers could look to
for this type of information. These include current customer
characteristics, competitor customer characteristics, market
research, and product strategy studies.
[0541] Exposure Recency Valuation
[0542] Some types of routine decisions follow a regular daily or
weekly pattern. The decisions are made over and over again. For
example, people often decide how and where they are going to eat
dinner right around dinner time. Many individuals frequently make
entertainment decisions just before the weekend. People decide what
to buy on Friday night as they browse through the store, and they
decide who to have repair the car on Saturday morning.
[0543] Some decisions are not periodic but they are still made at a
fixed time. Promoted television programs are aired at fixed times,
and most of the people who watch television programs make the
viewing decision just before the program is aired. Many people
appear to make their decision about whether or not to watch movies
about the time they first open in theaters. Public holidays are
fixed in time, and many purchasing decisions associated with these
occasions are made within a relatively narrow time window.
[0544] Under these types of market conditions, an advertising
exposure just before or at the point of decision is ideal.
According to most studies on learning theory, the more time that
elapses between an advertising exposure and the time of decision,
the less influence the advertising exposure will have at the time
of the decision. The value of an advertising exposure depends on
the amount of time that has elapsed between the time of exposure
and the time of decision. A measure of the value of an exposure
based solely on this time difference is termed exposure recency
value.
[0545] If, under certain market and product conditions, the rate of
influence decay is fast (see FIG. 30), then the value of a given
exposure is lost quickly. If the rate of decline is relatively
slow, then the exposure recency value will decline more slowly, and
exposure recency will not be as significant factor in determining
the value of a plan or schedule change.
[0546] Exposure recency values can be assigned using a table
similar to the one shown in FIG. 40. The table entries could depend
on a variety of factors related to the type of product being
advertised. The assignments shown in the table of FIG. 40 for
example, indicate that an advertisement aired on the same day as
the associated point of decision has full index value (i.e., 1).
Advertisements aired one day before are worth 60% of the value of
same-day advertisements, etc. These values may be determined
empirically, or may be based on other factors as determined by the
media planner. As before with other index values, more entries can
be added to increase the level of granularity.
[0547] All individual exposures for a given advertising spot have
the same exposure recency value. The value does not vary from one
audience member to another as it does for exposure or audience
valuation. Referring to the general scoring equation, if an
advertising spot is anytime on Wednesday, for example, and the
ideal advertising spot day is specified as Thursday, then the
exposure recency value for the spot is V.sub.T(a)=0.6 for all
members of the audience who are exposed to the advertisement on
that day.
[0548] If exposure recency is determined by measuring the number of
hours, rather than days, between the exposure and the time of
decision, as might be the case with snack food, which is often
consumed at mid-afternoon or during television prime-time, then a
table similar to with "Hours to Decision" would be used to assign
exposure recency index values on an hourly basis. Similarly, weeks
or other time periods might be specified.
[0549] Response Index
[0550] There are a number of factors associated with advertising
spots which could influence the level of response that audience
members will give advertising messages. These factors are unrelated
to the issues associated with the other index values already
discussed, such as exposure value, demographics, recency, or
advertisement quality. These factors are more closely related to
the host media and include a wide variety of information.
[0551] For example, the viewers of some television programs may, on
average, buy particular products more frequently than viewers of
other programs. In addition, some television programs may be more
effective in capturing and holding the attention of audience
members. Once captured, the audience may be less inclined to be
distracted during advertisements. This effect is known as program
inertia. Another factor may be the thematic nature of the media
message. As explained in the background section above, the theme or
subject material for some television programs or magazines may be
considered to be more consistent with some particular advertised
products. As another example, a certain program may be considered
to be unusually effective in setting the tone for a product, or an
advertiser may believe that audience members during certain times
of the day may be more (or less) attentive to advertisements.
[0552] The response index is defined as a composite of the various
types of factors listed above. In general, each of these factors
can be assigned to one of two categories, either host media
characteristics (such as the television program), or the type of
product being advertised (such as tooth paste). The response index
is not dependent on factors which are accounted for in other
indices such as demographic characteristics (this is included in
audience valuation), program loyalty levels for series programs
(this is included in multiple exposure valuation), audience skew
toward demographic group or another (this is included in
person-by-person analysis).
[0553] Fortunately, a number of sources of data are available which
can be used to derive a response index for product/media
combinations. Services such as MRI and Simmons assemble data that
relate media usage to product and service usage. These types of
data are widely available to advertisers and broadcasters, and
provide an excellent starting point from which response indices can
be estimated. It is anticipated that as additional product/service
usage information becomes available, the techniques for measuring
the relationship between the target audience and the
product/services will only improve.
[0554] In the broadest sense, the response index values required to
optimize an advertising plan or schedule are simply index numbers
associated with each of the spots in the plan. If, for example, a
simple advertising plan has only five spots, then the index values
might be assigned as follows:
1 Program Response Index Program A 0.85 Program B 1.00 Program C
0.78 Program D 1.15 Program E 1.65
[0555] As illustrated above, Program B, with a response index of
1.00, is arbitrarily selected as a base against which the response
index of other programs will be indexed. Program A, then is
considered to be 85% as effective as program B in persuading
audience members to purchase a particular advertised product.
Programs D and E, by contrast, have been determined to be superior
to Program B in influencing purchasing decisions. The average index
does not necessarily need to equal 1.0. The values just need to be
relatively consistent.
[0556] It should be noted that, using the methods of the present
invention, inferred relative response index values for spots in an
advertising plan can be computed. This is accomplished by supplying
the other index values for audience valuation, cost, recency, etc.
for the plan, and then computing the response index for each of the
spots in the plan which would be required to make all spots in the
plan of equal value. A comparison of these computed index values
could be used to identify inconsistencies in value for the spots in
the plan.
[0557] In assembling an advertising plan or schedule, a situation
may arise such as the following. The CPM for a particular program,
program A, is significantly higher than for other similar programs,
such as program B. Program A is very popular, and probably has
greater reach than program B, but the audience demographics may be
broader than required for the product to be advertised. Although
Program A seems to have a very loyal audience, this may push the
frequency levels higher than we require given the plan or schedule
that has already been developed. Finally, and possibly most
important, the programming for program A is more consistent with
the product than is the programming for program B. If the product
is advertised on program A rather than program B, it is necessary
to calculate how much of the added cost is going to the greater
reach and loyal audience in the context of our advertising plan or
schedule, and how much can be attributed to the program content. In
other words, how high is the response index of program A compared
to program B?
[0558] This question can be easily answered using the methods of
the present invention. Assume that accurate values for the response
index for both programs are included in equation (1) and that the
resulting scores are equal, meaning that if all things were
considered, there is no clear decision as to which program to
select. Thus:
S.sub.b(A)=S.sub.b(B)
[0559] or 26 i = 1 N A [ V I n ( i ) .times. d = 1 D V A d ( i ) ]
.times. V T ( A ) .times. V R ( A ) V C ( A ) 27 = i = 1 N B [ V I
n ( i ) .times. d = 1 D V A d ( i ) ] .times. V T ( B ) .times. V R
( B ) V C ( B )
[0560] Rearranging the terms and arbitrarily assuming that the
response index for program B is 1.0, then the equation becomes: 28
V R ( A ) = i = 1 N B [ V I n ( i ) .times. d = 1 D V A d ( i ) ]
.times. V T ( B ) V C ( B ) i = 1 N A [ V I n ( i ) .times. d = 1 D
V A d ( i ) ] .times. V T ( A ) V C ( A )
[0561] Thus, the ratio of the scores without a response index term
in either score equals the assumed response index of program A. If,
for example, the score S.sub.b(a) is computed for both programs
using the same base plan or schedule, and the score for program A
is 1000 and for program B is 2000 (in large part because of a large
difference in values for V.sub.C(a), the advertising cost for the
two programs), then it must be assumed that, all other things being
equal, people watching program A are at least twice as responsive
to advertising as they would be watching program B. If this is
true, then the obvious choice is to select program A. Otherwise
program B is selected.
[0562] One possible strategy to deal with uncertainty in response
indices is to optimize an advertising plan or schedule in multiple
phases. Each phase would focus on modifying the plan or schedule
for programs which are judged to have similar response indices.
[0563] Cost
[0564] The cost index V.sub.C(a) in equation (1) is the total
advertising cost, not cost per thousand (CPM). Most media planners
for network television do not always have convenient access to
absolute costs. They normally deal with costs in terms of CPMs.
This does not present a significant problem, however. A preferred
embodiment of a system which implements the methods of the present
invention will have access to person-by-person data. With this data
available, conversion to an index for absolute costs from CPM
values would not be difficult.
COMPREHENSIVE EXAMPLE
[0565] It is now possible to bring all the pieces of equation (1)
together into one comprehensive example. The starting point is the
base plan or schedule and exposure records which was introduced in
FIG. 35. The base plan or schedule includes six advertising spots.
The decision has been made to add one more spot at one of three
alternative times available on day 4.
[0566] The first step is to assemble the data shown in the table in
FIG. 41. This table includes four sections: the plan or schedule
with the exposure indicators for the 20 audience members, exposure
valuation indices, audience valuation indices, and subtotals for
each individual.
[0567] Exposure Valuation
[0568] Two valuation techniques are used to computing the exposure
values: weighted effective frequency, and time weighted effective
frequency. Scores are computed using both techniques so that the
differences in the results can be compared. Normally, a media
planner would select only one of the techniques for use in
optimizing an advertising plan or schedule. However, in certain
circumstances, it may be beneficial to use both techniques in
combination to optimize an advertising schedule or plan.
[0569] The frequency values for weighted effective frequency come
from the table shown in FIG. 23. For example, the exposure for
alternative B would be the second exposure of the plan or schedule
for person #1. According to the table shown in FIG. 23, exposure
number 2 has an index value of 0.8. This value is entered in for
alternative B, person #1. This person was not exposed to the
advertisements for alternative A or C, so no value is in those
positions under weighted effective frequency.
[0570] The values for time weighted effective frequency are the
individual audience member totals for each of the three
alternatives from the table shown in FIG. 35.
[0571] Audience Valuation
[0572] For this optimization example, only four demographic
characteristics are considered: age, gender, income, and county
size. Listed in the table shown in FIG. 41. are the details for
each of these characteristics for each person. Adjacent to each
demographic value is the index value for each characteristic.
Person #1, for example, is a 23 year old male. According to the
table shown in FIG. 37, a male of this age has no value to the
advertiser. He lives in a household with $16 K annual income. The
index value from the table shown in FIG. 38 for this level of
income is 0.5. He lives in an A sized county, which, according to
the table shown in FIG. 39, has an index value of 1.0 to the
advertiser. Similar entries are made for the other members of the
audience.
[0573] Subtotals
[0574] Next, each of the exposure values is multiplied with the
demographic index values for each person. These values are listed
as individual subtotals. This subtotal for alternative B for person
#3 using time weighted effective frequency, for example, is: 29 V I
n ( i ) .times. d = 1 D V A d ( i ) = 221 .times. 0.7 .times. 1.0
.times. 0.8 = 124
[0575] These individual values are then summed at the bottom of the
table shown in FIG. 40 according to the equation: 30 i = 1 N [ V I
n ( i ) .times. d = 1 D V A d ( i ) ]
[0576] Now, assume that the remaining indices necessary for
computing the scores are those shown in the table of FIG. 42. The
time recency indices are all 1.0 because all three program
alternatives are on the same night. The response indices for the
three alternatives are set by the advertiser based on information
from a variety of sources including market research, trade studies,
and/or judgment. The cost indices are derived from media rate
cards, and estimates. The base values for each of these indices are
arbitrary. Resetting all the time recency indices to 2.0, for
example, does not change the final ranking of the alternatives.
[0577] Results
[0578] The scores for each of the three alternatives for both
weighted effective frequency, and time weighted effective frequency
are listed in the table shown in FIG. 42. These are the values used
to decide which alternative to add to the base advertising plan or
schedule. The media objectives, assumptions about audience behavior
and learning, the values that placed on the various demographic
characteristics, and the actual exposures for the target audience
are all reflected in these values.
[0579] Weighted effective frequency, which, in this case,
emphasizes exposing audience members about three times, favors
alternative C, while time weighted effective frequency alternative
B. For purposes of comparison, it is interesting to note what the
decision may have been using one of the simpler decision rules,
previously explained above.
[0580] Average Frequency
[0581] All three proposed alternatives are identical. Each exposes
exactly five people, so the increase in average frequency is the
same. Even though the exposure numbers for each audience member is
different, the overall frequency for the audience doesn't
change.
[0582] Reach
[0583] All three alternatives are identical. None of the
alternatives exposes any audience member for the first time. Since
there is no change in the number of audience members exposed to the
advertisement, there is no change in the frequency rating.
[0584] Effective Frequency
[0585] If the presumed level of effectiveness is three exposures,
then all three alternatives exposes two people for the third time.
Again, all three choices are the same and no decision can be made
on this factor alone.
[0586] As illustrated by this example, based on the commonly-used
decision rules, then, it would be concluded that the three
alternatives are equally effective, when in fact, from a variety of
perspectives, they are not.
[0587] The viewing patterns for this example of 20 people were
specifically selected so that there would be no obvious patterns of
viewing, and so that there would be identical reach and frequency.
This was done to highlight the effectiveness that this integrated
method has in isolating important differences in scheduling
alternatives. When actual media-related access data is used, the
important differences between various options in advertising plans
or schedule are more pronounced, and the benefits to using this
method are even more significant.
[0588] Local Maximum
[0589] It is possible that the optimization process could converge
on a point which is the maximum score for small changes in a base
plan or schedule, but which is not the maximum score if all
possible plans or schedule for a time period were checked. If, for
example, an advertising plan or schedule was created by starting
with a base plan or schedule with no spots, and one spot after
another were added, it is still possible to arrive at a plan or
schedule which is less than optimal.
[0590] Suppose there is one person who is available to be exposed
to advertisements on any of 28 consecutive days. The goal is to
create an advertising plan or schedule which consists of a single
advertisement for this person which is optimal in the sense that it
maximizes average influence. As indicated in the table shown in
FIG. 43, the optimal position for a single advertisement would be
on the first day of the four week period. This single advertisement
would result in an average influence index of 13.5%. The "1 spot"
plot in FIG. 44 shows the level of influence that this one
advertisement would have over the four week period. If, assuming
another advertising spot was added to the first using the
optimization techniques of the present invention without changing
the timing of the first advertising spot, then this second
advertising spot should be on day 9. This would result in an
average influence of 24.3% for the four week period. If, to
continue the example, a third advertising spot was added to the
plan or schedule, assuming the first spot remains on day 1 and the
second on day 9, then the optimal position for the third spot would
be on day 15. This would result in an average influence of
32.8%.
[0591] Now, suppose the media planner is allowed to adjust the
position of spots numbers 2 and 3 in an effort to increase the
average influence. It is discovered that by moving spot 2 from day
9 to day 6, and spot 3 from day 15 to day 14, then the average
influence increases to 33.1%.
[0592] Plan or schedule number 3 is locally optimal in the sense it
solves the specific problem of maximizing influence with the
constraint that we cannot move the other spots. But plan or
schedule number 4 is globally optimal because there are no other
plans or schedule consisting of three spots which would produce a
higher average influence. However, finding the globally optimal
schedule required the adjustment of the timing of both spot 2 and
spot 3.
[0593] If a media planner is willing to change multiple spots at a
time, then the potential for finding a more optimal solution
improves. In order to be assured of finding the globally optimum
plan or schedule, it is necessary to test all possible alternatives
and combinations. If the goal is to optimize a plan or schedule of
r spots over a period of m days which had n available spots, then
the number of possible alternatives which would need to be tested
would be: 31 C r n = n ! r ! ( n - r ) !
[0594] If, for example, a media planner wanted to search all
possible options for an n=10 spot plan or schedule from among r=150
available spots, it would be necessary to compute the score for
about 10.sup.15 alternatives. If each computation required 1
second, the total computation would require about 37 million years,
which is impractical. The method described above, which allows the
movement of only one spot at a time, is probably adequate for most
purposes. It is important to note that various statistical and
computing techniques exist which overcome some of the problems
associated with identifying locally optimal solutions rather than
globally optimal ones. The methods of the present invention can be
easily adapted to incorporate those techniques.
[0595] User Interface
[0596] One feature of this integrated optimization method which
makes it particularly appealing is the simplicity of the user
interface. Despite the complexity of the internal processing, the
user interface is surprisingly simple. A media planner would
generally have most of the information required to use the system.
All that would be required to start the optimization process would
be a base advertising plan or listing, a listing of possible
alternative spots to add or remove, media objective values (such as
those shown in the tables of FIG. 37 and 40) and, if using weighted
effective frequency, the weighting values, or, if using time
weighted effective frequency, appropriate values for .alpha. and
.beta..
[0597] To optimize a plan or schedule, a planner would initialize
the system by entering the objective and weighting values. Then,
the base plan or schedule and the list of alternative slots would
be entered. The user would start the system processing on the two
lists. When complete, the system would return a listing of the
alternative spots ranked according to score. The user would select
from among the alternatives, add the spot to the list, and, if
desired, create another list of possible modifications and continue
the iterative process until the desired optimal schedule is
achieved.
[0598] Media Analysis Ratios
[0599] Another important consideration for media planning in
today's market is the ability to analyze and compare the viability
of various media vehicles for accomplishing the goals of the media
planner. The use of ratios for purposes of analysis can be applied
in many situations, including measuring the effectiveness of
competing media-related vehicles. Ratio analysis has been
previously used in accounting and management functions where ratios
such as "quick ratio," "short-term debt ratio," and "price/earnings
ratio" are used as comparative analysis tools to compare/contrast
competing businesses.
[0600] A similar type of ratio analysis can be accomplished by
using DME database 126 and DME 127. By extracting pertinent
media-related access information from DME database 126, a media
planner has some mechanism to compare and contrast the results
achieved by various competing alternative media vehicle. Subsets of
the data contained in DME database 126 can be selected to represent
certain programs, groups of programs, certain time slots, groups of
time slots, etc.
[0601] With this kind of flexible access to person-by-person
media-related access data, complete classes or families of analysis
can be defined which describes the exposure patterns to the media
for selected demographic groups. These classes may be composed of
collections of related exposure ratios which are composed of
various media-related exposure values. These collections of ratios,
then, can be used to analyze media offerings (such as related
television programs or time periods) in much the same way that
ratio analysis is used to measure the relative health of similar
businesses.
[0602] For example, given A and B, each of which can be either: a
television program; a time slot; a collection of television
programs; or a collection of time slots; various informative
media-related exposure values can be determined. This includes
determining the number of individuals in the sample audience who
were exposed to: some of A and some of B; all of A and all of B;
some of B and none of A; all of A and none of B; neither A nor B; A
and any during B including B; B and any during A; any during A and
any during B; A and any during B except B; etc.
[0603] After calculating the various exposure values identified
above, a ratio can be formed by using one of the exposure values as
the numerator and one of the exposure values as the denominator.
These various ratios can then be compared to ratios for a similar
selection of programs (or time slots) D and E.
[0604] Referring now to FIGS. 45 and 46, various examples of the
types and constitution of possible ratios are illustrated. As shown
in the table depicted in FIG. 45, depending on the type of analysis
to be performed, certain time segments are selected to represent
the audience viewing patterns at the desired point in time.
Similarly, the table in FIG. 46 describes several possible
combinations of media-related exposure values which will yield
information regarding audience viewing patterns. It should be noted
that the examples shown are for illustration purposes only. The
examples presented herein are representative, not exhaustive, and
those skilled in the art will readily recognize that many
variations are possible, depending on the type of information
desired.
SUMMARY
[0605] Television programming is a zero-sum game. If one
broadcasters gains a viewer, it typically means than another
broadcaster has lost one. Advertising plan or schedule optimization
is not zero-sum. An audience member could be far more valuable to
one advertiser than to another even though both advertisers are
pursuing identical demographic groups. This is because exposure
frequency and exposure timing are dependent on individual viewing
history for each advertisement, and the frequency curves for
scoring plans or schedule could be different for every advertiser
and every product. It is conceivable that there exists a globally
optimal advertising plan or schedule which spans all advertisers
and which uses all available advertising spots of all broadcasters
at peak efficiency. There would still be audience members who are
under or over exposed to particular advertising messages, but under
globally optimal conditions, there would be no way to rearrange
advertisements in such a way that the total score for all
advertisers increases.
[0606] There is, of course, no compelling reason on the part of
broadcasters or advertisers to seek a global optimum. The point,
however, is that if one advertiser optimizes a plan or schedule, it
does not necessarily reduce the flexibility that other advertisers
have in optimizing their plans or schedule. In fact, it may
actually improve the available mix of spots that would be available
to other advertisers. This would be desirable for both broadcasters
and advertisers.
[0607] The task of optimizing an advertising plan or schedule where
the vehicle options number less than a dozen, as is the case with
network television, is significant. The task of optimizing a plan
or schedule where the vehicle options number in the thousands is
insurmountable without the systematic techniques and methods as
described in this specification. The ability to gather
media-related exposure data improves as electronic media becomes
more popular. In the future, it is likely that data for larger
sample sizes will be available, the data will be cheaper to
accumulate, and more data will be available. This would further
suggest the need for better mechanisms for analyzing the data as
described herein. By using the methods of the present invention, a
media planner can effectively distribute advertisements over time
or space based on previous or anticipated individual or collective
advertising exposures.
[0608] While the invention has been particularly shown and
described with reference to preferred embodiments thereof, it will
be understood by those skilled in the art that various changes in
form and details may be made therein without departing from the
spirit and scope of the invention.
* * * * *