U.S. patent application number 11/402354 was filed with the patent office on 2006-08-17 for dynamically optimizing the presentation of advertising messages.
Invention is credited to Scott Eric Lipsky, Chen Yu.
Application Number | 20060184421 11/402354 |
Document ID | / |
Family ID | 26862823 |
Filed Date | 2006-08-17 |
United States Patent
Application |
20060184421 |
Kind Code |
A1 |
Lipsky; Scott Eric ; et
al. |
August 17, 2006 |
Dynamically optimizing the presentation of advertising messages
Abstract
A facility for adjusting the execution of an advertising
campaign in which advertising messages are presented using a
plurality of advertising alternatives as described. During a first
time period, the facility presents advertising messages using each
of the advertising alternatives in accordance with an initial
allocation for each of the advertising alternatives. Also during
the first time period, the facility tracks the performance of the
advertising campaign with respect to each of the advertising
alternatives. Based upon the tracking during the first time period,
the facility attributes a performance score to each of the
advertising alternatives for the first time period. The facility
compares these scores, and, based upon the comparison, adjusts the
allocations for the advertising alternatives so as to increase one
or more allocations for advertising alternatives which compare
favorably in the comparison, and so as to reduce one or more
allocations for advertising alternatives comparing disfavorably in
the comparison. The facility then, during a second time period,
presents advertising messages using each of the advertising
alternatives in accordance with the adjusted allocation for each of
the advertising alternatives.
Inventors: |
Lipsky; Scott Eric;
(Seattle, WA) ; Yu; Chen; (Pittsburgh,
PA) |
Correspondence
Address: |
PERKINS COIE LLP;PATENT-SEA
P.O. BOX 1247
SEATTLE
WA
98111-1247
US
|
Family ID: |
26862823 |
Appl. No.: |
11/402354 |
Filed: |
April 12, 2006 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09626266 |
Jul 25, 2000 |
7031932 |
|
|
11402354 |
Apr 12, 2006 |
|
|
|
60167055 |
Nov 22, 1999 |
|
|
|
Current U.S.
Class: |
705/14.42 |
Current CPC
Class: |
G06Q 30/0277 20130101;
G06Q 30/0273 20130101; G06Q 30/02 20130101; G06Q 30/0242 20130101;
G06Q 30/0243 20130101 |
Class at
Publication: |
705/014 |
International
Class: |
G06Q 30/00 20060101
G06Q030/00 |
Claims
1-28. (canceled)
29. A computer-readable medium whose contents cause a computing
device to adjust the execution of an advertising campaign for
presenting advertising messages to a plurality of users, the
advertising campaign having a plurality of advertising alternatives
for presenting advertising messages, by: during a first temporal
range, presenting advertising messages to users among the plurality
using each of the advertising alternatives in accordance with an
initial allocation for each of the advertising alternatives;
tracking the performance of the advertising campaign with respect
to each of the advertising alternatives across the plurality of
users; based upon the tracking during the first temporal range,
attributing a performance score to each of the advertising
alternatives for the first temporal range; comparing the scores
attributed to the advertising alternatives for the first temporal
range, wherein the comparison is performed using confidence
intervals about the performance scores; based upon the comparison,
adjusting the allocations for the advertising alternatives so as to
increase one or more allocations for advertising alternatives
comparing favorably in the comparison and so as to reduce one or
more allocations for advertising alternatives comparing
disfavorably in the comparison; and during a second temporal range,
presenting advertising messages using each of the advertising
alternatives in accordance with the adjusted allocation for each of
the advertising alternatives.
30. The computer-readable medium of claim 29 wherein the
performance score attributed to each advertising alternative
reflects the percentage of advertising messages presented using the
advertising alternative that were selected by users to which they
were presented.
31. The computer-readable medium of claim 31 wherein the
performance score attributed to each advertising alternative
reflects the percentage of advertising messages presented using the
advertising alternative that were selected by users to which they
were presented to access additional information relating to an
advertiser associated with the campaign where the user performed a
further predefined action with respect to the additional
information.
32. The computer-readable medium of claim 31 wherein the further
predetermined action is transacting a purchase.
33. The computer-readable medium of claim 31 wherein the further
predetermined action is registering as a user.
34. The computer-readable medium of claim 31 wherein the further
predetermined action is specified by the advertiser.
35. The computer-readable medium of claim 29 wherein the
performance score attributed to each advertising alternative
reflects the total amount of purchases transacted as a result of
presentations using the advertising alternative.
36. The computer-readable medium of claim 29 wherein the
performance score attributed to each advertising alternative
reflects the total amount of purchases transacted as a result of
presentations using the advertising alternative as compared to the
number of advertising messages presented using the advertising
alternative.
37. The computer-readable medium of claim 29 wherein the
performance score attributed to each advertising alternative
reflects the cost of presenting advertising messages using the
advertising alternative.
38. The computer-readable medium of claim 37 wherein the reflected
cost of presenting advertising messages using the advertising
alternative includes the cost of purchasing advertising space in
which to present advertising messages using the advertising
alternative.
39. The computer-readable medium of claim 37 wherein the reflected
cost of presenting advertising messages using the advertising
alternative includes the cost of delivering data comprising
advertising messages using the advertising alternative.
40. The computer-readable medium of claim 29 wherein the contents
of the computer-readable medium further cause the computing device
to: during the second temporal range, tracking the performance of
the advertising campaign with respect to each of the advertising
alternatives; based upon the tracking during the first temporal
range, attributing a performance score to each of the advertising
alternatives for the first temporal range; comparing the scores
attributed to the advertising alternatives for the second temporal
range; based upon the comparison of scores attributed to the
advertising alternatives for the second temporal range, again
adjusting the allocations for the advertising alternatives so as to
increase one or more allocations for advertising alternatives
comparing favorably in the comparison and so as to reduce one or
more allocations for advertising alternatives comparing
disfavorably in the comparison; and during a third temporal range,
presenting advertising messages using each of the advertising
alternatives in accordance with the twice-adjusted allocation for
each of the advertising alternatives.
41. The computer-readable medium of claim 29 repeated for a
plurality of iterations over a plurality of further temporal
ranges.
42. A computer system for adjusting the execution of an advertising
campaign for presenting advertising messages, the advertising
campaign having a plurality of advertising alternatives for
presenting advertising messages, comprising: a presenting component
that presents advertising messages using each of the advertising
alternatives in accordance with an allocation for each of the
advertising alternatives, the presenting component presenting
advertising messages using each of the advertising alternatives in
accordance with an initial allocation for each of the advertising
alternatives during a first time period, the presenting component
presenting advertising messages using each of the advertising
alternatives in accordance with an adjusted allocation for each of
the advertising alternatives during a second time period; a
tracking component that tracks the performance of the advertising
campaign with respect to each of the advertising alternatives
during a first time period; a scoring component that, based upon
the tracking during the first time period, attributes a performance
score to each of the advertising alternatives for the first time
period; a comparison component that compares the scores attributed
to the advertising alternatives for the first time period, wherein
the comparison is performed using confidence intervals about the
performance scores; and an allocation adjustment component that,
based upon the comparison, adjusts the allocations for the
advertising alternatives so as to increase one or more allocations
for advertising alternatives comparing favorably in the comparison
and so as to reduce one or more allocations for advertising
alternatives comparing disfavorably in the comparison.
43. The computer system of claim 42, further comprising a cycling
component that triggers repetition of tracking, scoring, comparing,
and adjustment over a plurality of further time periods.
44-51. (canceled)
52. A computer-readable medium whose contents cause a computer
system to present advertising messages in a group of advertising
messages, by: during an evaluation period, presenting the
advertising messages of the group; assessing the effectiveness of
presenting each of the advertising messages during the evaluation
period; assigning presentation weights to the advertising messages
of the group in accordance with their assessed effectiveness; and
during a weighted presentation period, presenting the advertising
messages of the group with relative frequencies that are in
accordance with their weights.
53. The computer-readable medium of claim 52 wherein the weighted
presentation period is a second evaluation period, and wherein the
assessing and assigning is repeated for the second evaluation
period.
54. The computer-readable medium of claim 52 wherein the
advertising messages are presented at a plurality of advertising
sites, and wherein the assessing assesses the effectiveness of
presenting each of the advertising messages at each of the
advertising sites, and wherein the assigning step assigns an
advertising message separate weights for each of the advertising
sites, and wherein the weights for the advertising site on which an
advertising message is to be presented are used to select the
advertising messages to be presented on that advertising site.
55. The computer-readable medium of claim 29 wherein the adjusting
adjusts the allocations such that all of the adjusted allocations
are nonzero, such that advertising messages are presented using
each of the advertising alternatives during the second time
period.
56. The computer-readable medium of claim 29 wherein each of the
plurality of advertising alternatives is a different advertising
message.
Description
CROSS REFERENCE TO RELATED APPLICATION
[0001] This application claims the benefit of provisional U.S.
Patent Application No. 60/167,055 filed Nov. 22, 1999 which is
hereby incorporated by reference.
TECHNICAL FIELD
[0002] The present invention is directed to Internet advertising
techniques.
BACKGROUND
[0003] The World Wide Web ("the Web") is a medium for making
content available to requesting users, also known as "publishing"
content. A Web publisher may provide content of various types,
including visual Web pages having textual and static visual
contents, as well as such other content such as animated images and
audio and video sequences. The relatively low cost of publishing
content on the Web enables many organizations and individuals to
act as Web publishers.
[0004] To access content made available on the Web, a user
typically directs a Web client program, or "browser," executing on
his or her computer system to obtain and display a particular unit
of content. Units of content are each identified by an address,
also called a Uniform Resource Locator, or URL. The user may direct
the browser to obtain and display a unit of content by directly
entering the URL for the unit of content, or by selecting a link or
bookmark with which the URL is associated.
[0005] Internet advertising is a practice in which a Web publisher
adds an advertising message or other content provided by a third
party advertiser to the content provided by the Web publisher in
exchange for payment or other consideration from the advertiser.
For example, a Web-based travel agency may pay the publisher of a
Web site directed to tips for travelers to add a banner
advertisement for the Web-based travel agency to some of the Web
pages comprising the Web site. Such a banner advertisement
typically presents visual information promoting the advertiser that
takes up a portion of the area of the Web page. Some banner
advertisements further constitute a link to a Web site associated
with the advertiser, so that the user may click anywhere within the
banner advertisement to display, or "click through to," that Web
site, and perform additional actions there, such as purchasing a
product or registering with the advertiser Web site. When such
additional actions are performed, it is said that a "conversion"
has occurred.
[0006] In the terminology of Web advertising, a particular banner
advertisement may be called a "creative" or a "advertising
message." The different locations on Web pages of publisher Web
sites that the publisher makes available for advertising are called
"placements." An instance of presenting a particular creative in a
particular placement for a particular use is called an
"impression." When a publisher enters into an advertising agreement
with an advertiser in which the advertiser purchases for a single
price the opportunity to display a certain number of impressions in
each of a number of placements, these placements and impressions
are said to constitute a "cost package."
[0007] As computer use, and particularly the use of the Web,
becomes more and more prevalent, the volume of Internet advertising
presented grows larger and larger. As this volume continues to
increase, the need to optimize the effectiveness of advertising
increases significantly.
[0008] In this connection, it is now common for an advertiser to
mount a campaign for an advertiser's cause, in which it presents a
number of different "alternative" advertising messages for the
cause in each of a number of different placements. Indeed, many
advertising campaigns utilize multiple cost packages of placements
purchased from multiple publishers.
[0009] Such an extensive campaign can be difficult to manage in an
active manner. In particular, it can be difficult to determine
whether to increase or decrease the number of impressions presented
in a particular cost package, whether to increase or decrease the
number of impressions presented in a particular placement, and
whether to increase or decrease the rate at which a particular
creative is presented.
[0010] Accordingly, a facility for automatically and dynamically
optimizing the use of resources in an advertising campaign would
have significant utility.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] FIG. 1 is a high-level block diagram showing the environment
in which the facility preferably operates.
[0012] FIG. 2 is a flow diagram showing the steps preferably
performed by the facility in order to optimize the performance of
an advertising campaign.
[0013] FIGS. 3-5 are data structure diagrams showing the analysis
of sample data by the facility.
DETAILED DESCRIPTION
[0014] A software facility for automatically and dynamically
optimizing the use of resources in an advertising campaign is
provided. In designing an advertising campaign, an advertising
service selects cost packages to use in presenting advertising
messages for the campaign, each of which constitutes an opportunity
to present a certain number advertising messages in each one or
more placements. The advertising service further selects a number
of advertising messages to present as part of the campaign. The
facility conducts the campaigns, presenting the selected
advertising messages in the placements within the selected cost
packages While conducting the campaign, facility maintains
statistics indicating the level of performance of each of the
advertising messages, placements, and cost packages.
[0015] After the campaign has been conducted for a period of time,
the facility compares the performance of the selected advertising
messages, placements, and cost packages, and scores their relative
performance. Performance scores for these aspects of the campaign,
sometimes called "advertising alternatives," may be based upon a
variety of factors, including click-throughs, conversions, and
sales produced by these advertising alternatives, as well as their
cost. To the extent that reallocations between advertising
messages, placements, and cost packages are possible, the facility
performs reallocations from advertising messages, placements, and
cost packages indicated by their performance scores to be
performing at a low level to higher-performing ones.
[0016] For example, reallocating between cost packages may involve
negotiating with the publisher or other seller of a
higher-performing cost package to increase the volume of the
higher-performing cost package, as well as negotiating with the
publisher or other seller of a lower-performing cost package to
cancel or decrease the volume of the lower-performing cost package.
Reallocating between the placements of a cost package may involve
negotiating with the publisher or other seller of the cost package
to increase the volume of the higher-performing allocations of the
cost package and decrease the volume of the lower-performing
allocations of the cost package. Reallocating between advertising
messages presented in a placement may involve increasing the
probability that higher-performing advertising messages are served
in response to an advertising message request for the placement and
decreasing that probability for lower-performing advertising
messages. After adjusting these allocations in accordance with the
effectiveness scores, the facility continues the campaign using
these new allocations, again maintaining performance statistics in
order to later perform further reallocations.
[0017] In some embodiments, the facility manages the operation of
several advertising campaigns simultaneously. For example, the
facility may simultaneously manage the operation of campaigns for
several different advertisers. This further extends the facility's
ability to optimize campaign performance, by enabling the facility
to exchange measures of advertising alternatives between campaigns,
thereby allowing reallocation of advertising alternatives that
otherwise could not be reallocated, e.g., allowing reallocation
between two cost packages whose seller declines to renegotiate.
[0018] By automatically optimizing the operating of advertising
campaigns on-the-fly based upon the actual performance of the
various advertising alternatives that comprise the campaign in this
manner, the facility greatly increases the effectiveness of the
campaign and produces significant efficiencies in operating the
campaign.
[0019] FIG. 1 is a high-level block diagram showing the environment
in which the facility preferably operates. The diagram shows client
computer systems 111 and 112. An Internet user preferably uses one
of these client computer systems to connect, via the Internet 100,
to an Internet publisher computer system, such as Internet
publisher computer systems 130 and 140, to retrieve and display a
Web page.
[0020] In cases where an Internet advertiser, through the Internet
advertising service, has purchased advertising space on the Web
page provided to the Internet user computer system by the Internet
publisher computer system, the Web page contains a reference to a
URL in the domain of the Internet advertising service computer
system 120. When a user computer system receives a Web page that
contains such a reference, the Internet user computer systems sends
a request to the Internet advertising service computer system to
return data comprising an advertising message, such as a banner
advertising message. When the Internet advertising service computer
system receives such a request, it selects an advertising message
to transmit to the Internet user computer system in response the
request, and either itself transmits the selected advertising
message or redirects the request containing an identification of
the selected advertising message to an Internet content distributor
computer system, such as Internet content distributor computer
systems 170 and 180. When the Internet user computer system
receives the selected advertising message, the Internet user
computer system displays it within the Web page.
[0021] The displayed advertising message preferably includes one or
more links to Web pages of the Internet advertiser's Web site. When
the Internet user selects one of these links in the advertising
message, the Internet user computer system dereferences the link to
retrieve the Web page from the appropriate Internet advertiser
computer system, such as Internet advertiser computer system 150 or
160. In visiting the Internet advertiser's Web site, the Internet
user may traverse several pages, and may take such actions as
purchasing an item, bidding in an auction, or registering as a user
of the Internet advertiser. Revenue from such actions typically
finances, and is often the motivation for, the Internet
advertiser's Internet advertising.
[0022] The Internet advertising service computer system 120
preferably includes one or more central processing units (CPUs) 121
for executing computer programs such as the facility, a computer
memory 122 for storing programs and data while they are being used,
a persistent storage device 123 such as a hard drive for
persistently storing programs and data, and a computer-readable
media drive 143, such as a CD-ROM drive, for reading programs and
data stored on a computer-readable medium.
[0023] While preferred embodiments are described in terms of the
environment described above, those skilled in the art will
appreciate that the facility may be implemented in a variety of
other environments, including a single, monolithic computer system,
as well as various other combinations of computer systems or
similar devices.
[0024] In an embodiment discussed in detail below, the facility
judges the effectiveness of advertising messages, placements, and
cost packages in terms of the conversion rates that they produce.
In some embodiments, however, the facility preferably uses a rating
function that rates the effectiveness of each alternative
advertising message when presenting each placement and cost package
in which it is presented based upon other factors. The rating
function preferably has a positive reaction, for a particular
advertising message and publisher, to one or more of the following
factors: (1) "conversion rate," the percentage of presentations of
the advertising message at the publisher to users that coincided
with users that performed some action on the advertiser's web site,
such as visiting it or placing an order; (2) "click rate," the
percentage of times the advertising message was presented at the
publisher and was clicked on by the user viewing the advertising
message; and/or (3) "average transaction value," the average dollar
amount of orders placed by customers that viewed the advertising
message at the publisher. The rating function also preferably has a
negative reaction, for a particular advertising message and
publisher, to the financial cost of presenting the advertising
message at the publisher. After the rating function is applied to
each advertising message for a particular publisher to produce raw
rating scores, the raw rating scores are preferably normalized so
that they sum to 100% for the publisher.
[0025] In various embodiments, the rating function is applied to
update the weights at various frequencies, including frequencies
such as once per week, once per day, once per hour, or
continuously.
[0026] In certain embodiments, the rating function limits the
extent to which the rating of a particular advertising message and
publisher can change during a single period. For example, in one
preferred embodiment, the rating function is limited to changing no
more than 5% per period.
[0027] In some embodiments, the facility retains the initial
weightings without applying the rating function until a minimum
number of advertising message presentations is reached. For
example, the facility may defer re-weighting until a minimum number
of total presentations of any of the advertising messages within
the group at any of the publishers is reached. In additional
embodiments, re-weighting is deferred until each advertising
message, each publisher, or each combination of an advertising
message and a publisher has a certain number of presentations.
[0028] In one embodiment, the facility maintains a history of the
advertising presentation weights used during each period to
facilitate the review of the performance of the advertising
messages using different weights. In this manner, the facility can
analyze the effects of particular weighting changes on
performance.
[0029] FIG. 2 is a flow diagram showing the steps preferably
performed by the facility in order to optimize the performance of
an advertising campaign. This process is preferably repeated by the
facility periodically for each advertising campaign that is being
managed by the facility. In steps 201-210, the facility loops
through each cost package of placements that has been selected for
the facility. In steps 202-206, the facility loops through each
placement of the current cost package. In step 203, the facility
generates scores measuring the effectiveness of advertising
messages that have been presented in the current placement. In a
preferred embodiment, effectiveness scores correspond to conversion
rates--that is, the fraction of impressions of the advertising
messages that lead to a conversion on the advertiser's web site,
such as a purchase. In this embodiment, the effectiveness score of
an advertising message is calculated by dividing the number of
conversions produced by presenting the advertising message in the
current placement by the number of times the advertising message
was presented in the current placement, also described as the
number of impressions of the advertising message in the current
placement. As is discussed elsewhere herein, additional embodiments
of the present invention utilize a variety of other formulae for
generating effectiveness scores, which are in these embodiments
utilized in step 203.
[0030] In step 203, the facility further compares the effectiveness
scores generated for the advertising messages being presented in
the current placement. In one preferred embodiment, the facility
compares the absolute effectiveness scores of the advertising
messages. In another preferred embodiment, the facility compares
confidence intervals about the scores. Confidence intervals relate
to statistical measures, such as effectiveness scores, that are
based on a finite number of observations, or "samples." A
confidence interval indicates a range of values around, or "about,"
the statistical measure into which there is a particular
probability that the same statistical measure based upon a much
larger population of samples will fall. In one embodiment, the
facility compares 80 percent confidence intervals about the scores,
meaning that the facility compares ranges about the scores into
which there is an 80 percent chance that scores based on a much
larger population of samples will fall.
[0031] Equations (1)-(12) below show the derivation of the formula
in equation (12) preferably used by the facility to calculate an 80
percent confidence interval about each effectiveness score.
Equation (1) indicates that the confidence interval
(a,b).sub..alpha. for a particular confidence level .alpha. is
defined as the statistical measure {overscore (x)}, here the
effectivess score.+-.a confidence interval radius c.sub..alpha. for
that confidence level. (a,b)=.sub..alpha.={overscore
(x)}.+-.c.sub..alpha. (1)
[0032] Equation (2) indicates that the confidence interval radius
c.sub..alpha. is equal to a z-score z.sub..alpha. for the
confidence level multiplied by the standard deviation of the larger
population .sigma. and divided by the square root of the number of
samples n upon which the statistical measure is based. c .alpha. =
z .alpha. .times. .sigma. n ( 2 ) ##EQU1##
[0033] Equation (3) indicates that, in accordance with the Central
Limit Theorem, the standard deviation .sigma. of the larger
population can be assumed to be equal to the standard deviation s
of the samples upon which the statistical measure is based.
.sigma.=s (3)
[0034] Equation (4) indicates that the standard deviation s of the
samples upon wich the statistical measure is based is equal to the
square root of the product of the proportion of successes p times
the proportion of failures (1-p). s = p .function. ( 1 - p ) ( 4 )
##EQU2##
[0035] Equation (5) shows the substitution of equations (3) and (4)
in equation (1). c .alpha. = z .alpha. .times. p .function. ( 1 - p
) n ( 5 ) ##EQU3##
[0036] Equation (6) shows that, in this case, the number of samples
n upon which the statistical measure is based is the number of
impressions in which the current advertising message was presented
in the current placement. n=impressions (6)
[0037] Equation (7) shows that the proportion of successes p is
equal to the number of conversions that resulted from presenting
the current advertising message in the placement divided by the
number of impressions. p = conversions impressions ( 7 )
##EQU4##
[0038] Equation (8) shows the substitution of equations (6) and (7)
in equation (5). c .alpha. = z .alpha. .times. conversions
impressions .times. ( 1 - conversions impressions ) impressions ( 8
) ##EQU5##
[0039] Equation (9) shows the simplification of equation (8). c
.alpha. = z .alpha. .times. conversions .function. ( impressions -
conversions ) impressions ( 9 ) ##EQU6##
[0040] Equation (10) shows that the z-score for an 80 percent
confidence interval is 1.28, as derived from statistical tables.
z.sub.80%=1.28 (10)
[0041] Equation (11) shows the substitution of equation (10) in
equation (9). c 80 .times. % = 1.28 .times. conversions .function.
( impressions - conversions ) impressions ( 11 ) ##EQU7##
[0042] Equation (12) shows the substitution of equation (11) in
equation (1), and indicates how to calculate an 80 percent
confidence interval about a conversion rate effectivess score for a
particular advertising message in a particular placement given the
number of conversions and the number of impressions measured for
that advertising message in that placement. ( a , b ) 80 .times. %
= .times. conversions impressions .+-. .times. 1.28 .times.
conversions .function. ( impressions - conversions ) impressions (
12 ) ##EQU8##
[0043] In step 204, the facility reallocates the relative weights
of the advertising messages for the current placement based upon
the comparison performed in step 203. When an advertising message
presentation request is received requesting an advertising message
to present in the current placement, these weights are used to
determine which advertising message to return. The weight for a
particular advertising message reflects the relative likelihood
that that advertising message will be returned in response to each
such request. A number of embodiments of the facility utilize
various reallocation schemes in step 204. In some embodiments, one
or more advertising messages are "dropped" by assigning them a
weight of zero, thereby increasing the relative weights of the
advertising message that are not dropped. In these embodiments, any
number of advertising messages may be dropped: a single
lowest-performing advertising messages, the bottom-performing n
advertising messages, the lowest-performing half of the advertising
messages, all but the highest-performing advertising message, etc.
In other embodiments, the facility incrementally adjusts the weight
of each advertising message, incrementally reducing the weights of
lower-performing advertising messages, and incrementally increasing
the weights of higher-performing advertising messages. Additional
reallocation schemes may also be employed.
[0044] In step 205, the facility forecasts the effectiveness scores
for the advertising messages based upon the new allocation of
advertising messages created in step 204. In step 206, the facility
loops back to step 202 to process the next placement in the current
cost package. After the last placement in the current cost package
has been processed, the facility continues in step 207. In step
207, the facility forecasts the effectiveness scores of the
placements in the current cost package based upon the effectiveness
scores forecasted in step 205 for the advertising messages
presented in those placements. In step 208, the facility compares
the effectiveness scores forecasted in step 207 for the placements
in the current cost package. Step 208 is preferably performed in a
manner similar to the comparison performed in step 203. In step
209, the facility reallocates the relative weights of the
placements in the cost package based upon the comparison of the
effectiveness scores forecasted for them in step 208. Step 209 is
preferably performed in a manner similar to that of step 204.
Reallocating between the placements of a cost package may involve
negotiating with the publisher or other seller of the cost package
to increase the volume of the higher-performing allocations of the
cost package and decrease the volume of the lower-performing
allocations of the cost package. In step 210, the facility loops
back to step 201 to process the next cost package. After the last
cost package is processed, the facility preferably continues in
step 211. In step 211, the facility forecasts cost effectiveness
scores for the cost packages based upon effectiveness scores
forecasted in step 207 for the placements in each cost package. In
step 212, the facility compares the cost effectiveness scores
forecasted in step 211 for the cost packages. Step 212 is
preferably performed in a manner similar to the comparison of step
203. In step 213, the facility reallocates the relative weights of
the cost packages based upon the comparison of step 212. Step 213
is preferably performed in a manner similar to the comparison of
step 204. Reallocating between cost packages may involve
negotiating with the publisher or other seller of a
higher-performing cost package to increase the volume of the
higher-performing cost package, as well as negotiating with the
publisher or other seller of a lower-performing cost package to
cancel or decrease the volume of the lower-performing cost package.
In step 214, the facility applies the reallocations performed in
steps 204, 209, and 213. After step 214, these steps conclude.
[0045] FIGS. 3-5 are data structure diagrams showing the analysis
of sample data by the facility. FIGS. 3-5 show sample data for a
single cost package. As is discussed herein, however, the facility
is preferably capable of managing campaigns that make use of more
than one cost package, and, indeed, managing several such campaigns
simultaneously.
[0046] FIG. 3 is a data structure diagram showing data relating to
the performance of a sample cost package. A table 300 contains rows
311-325. Rows 311-324 each correspond to a different combination of
a placement and an advertising message presented in that placement.
Row 325 contains values aggregated from rows 311-324. Each row
contains a field in each of the following columns: a placement
column 301 identifying a placement in the cost package; an
advertising message 302 identifying an advertising message
presented in that placement; an allocation column 303 showing the
current allocation weighting for the advertising message without
the placement; an impressions column 304 indicating the number of
times that the identified advertising message has been presented in
the identified placement; a conversions column 305, indicating the
number of such impressions that have resulted in a conversion event
at the advertiser's web site; a conversion rate column 306
indicating the quotient of conversions over impressions; a cost
column 307 indicating, in row 325, the cost of presenting the
counted impressions; and a cost per conversion column 308
indicating, in row 325 the ultimate cost incurred per conversion.
For example, the contents of row 312 show that advertising message
2 has been presented in placement 1 234,333 times and produced 234
conversions, thus achieving a conversion rate of 0.0010.+-.0.0404.
The conversion rate of 0.0010 is obtained by the facility by
dividing 234 conversions by 234,333 impressions. The confidence
interval rate is of 0.0404 is obtained in accordance with equation
(11) by subtracting the number of conversions, 23, from the number
of impressions, 243,655, then multiplying by the number of
conversions, 23, then taking the square root of this quantity, then
multiplying by 1.28 and dividing by the number of impressions,
234,655. Row 325 shows that using cost package 1 in the current
campaign has produced a total of 1,564 conversions, at a total cost
of $200,000, for an ultimate cost per conversion of $13.28. It can
be seen that advertising message 3 has the highest conversion rate
for placement 1, advertising message 1 has the highest conversion
rate for placement 2, advertising message 1 has the highest
conversion rate for placement 3, advertising message 2 has the
highest conversion rate for placement 4, and advertising message 2
has the highest conversion rate for placement 5.
[0047] FIG. 4 is a data structure diagram showing reallocation
among advertising messages for each placement. Table 400 shows the
result of reallocating weights among the advertising messages
displayed in each placement. By comparing the new allocation
weights shown in allocation column 403 to the conversion rates
shown in column 306 of FIG. 3, it can be seen that the facility has
assigned a new weight of 1 to the advertising message for each
placement having the highest conversion rate, and an allocation
weight of 0 to each of the other advertising messages within the
placement. For example, advertising message 3 having the highest
conversion rate within placement 1, is assigned an allocation
weight of 1for placement 1, while advertising messages 1 and 2 are
assigned an allocation weight of 0 for placement 1. Impressions
column 404 shows the number of impressions forecasted for each
combination of placement and advertising message in the next
analysis cycle. For placement and advertising message combinations
having an allocation weight of 0, 0 impressions are projected. For
placement and advertising message combinations having weight of 1,
a number of impressions equal to the sum of observed impressions
for all of the advertising messages for that placement is
projected. For example, the 721,000 impressions projected for
advertising message 3 in placement 1 is the sum of the number of
impressions observed for all advertising messages in placement 1.
Conversions column 405 contains a projection of the number of
conversions that will be produced by those impressions, and is
obtained by multiplying the projected number of impressions in
column 404 by the observed conversion rate shown in column 306. For
example, the projection of 313,526 conversions shown for placement
1 in advertising message 3 is the product of 721,000 impressions
and a conversion rate of 0.0188. The conversions field in row 425
is updated to be the sum of projected conversions. Similarly, the
cost per conversion field in row 425 is updated to a projection of
$4.30 per conversion by dividing the total cost of $200,000 by the
projected total number of conversions, 46,551. Thus, it can be seen
that, by reallocating advertising message weights as shown, the
facility projects that is will reduce an ultimate cost per
conversion from $13.28 to $4.30.
[0048] FIG. 5 is a data structure diagram showing the comparison of
the performance of placements within a cost package. Table 500
shows that the facility, for each placement, attributes the
conversion rate of the non-zero weighted advertising message to the
entire placement. On that basis, the facility compares the
conversion rates attributed to each placement for the purpose of
performing a reallocation among placements. The facility determines
that placement 3 has the highest conversion rate, 0.0233.
Accordingly, the facility, if possible, reweights the placements so
as to increase the allocation to placement 3 and decrease the
allocation to lower-performing placements, such as placements 4 and
5. Based on these reallocations among the placements of cost
package 1, the facility preferably repeats its forecasting to
determine a new cost per conversion for cost package 1. On this
basis, it preferably compares the cost per conversion projected for
cost package 1 to costs per conversion similarly projected for the
other cost packages of the campaign, and performs a reallocation
away from cost packages having a higher cost per conversion toward
cost packages having a lower cost per conversion.
[0049] It will be understood by those skilled in the art that the
above-described facility could be adapted or extended in various
ways. For example, while advertising campaigns in which the
advertising alternatives are advertising messages, placements, and
cost package are discussed above, the facility may be
straightforwardly adapted to manage advertising campaigns having
more, fewer, or different advertising alternatives. Additionally,
the facility may employ a variety of different types of
effectiveness scores for comparing the effectiveness of different
advertising alternatives. Further, in the foregoing, embodiments of
the facility have been described that automatically manage
advertising campaigns in which advertising messages are delivered
via the World Wide Web to users using general-purpose computer
systems executing web browsers. Additional embodiments of the
facility may be used to manage campaigns in which advertising
messages are presented to users via various other communication
channels and other types of devices, including special-purpose
devices such as personal digital assistants, cellular and satellite
telephones, devices installed in automobiles and other vehicles,
pagers, automatic teller machines, televisions, and other home
appliances. Moreover, a further embodiment of the facility is
adapted to allocate advertising alternatives separately for each of
a number of these alternative outlet types, so that, for example,
an advertisement that is particularly effective on a certain
publisher via personal digital assistants is presented more
frequently on personal digital assistants than alternative
advertisements.
[0050] While the foregoing description makes reference to various
preferred embodiments, the scope of the invention is defined solely
by the claims that follow and the elements recited therein.
* * * * *