U.S. patent application number 14/134606 was filed with the patent office on 2014-11-20 for methods and apparatus for optimizing advertisement allocation.
This patent application is currently assigned to TREMOR VIDEO, INC.. The applicant listed for this patent is Tremor Video, Inc.. Invention is credited to Steven Lee, Tadashi Yonezaki.
Application Number | 20140344048 14/134606 |
Document ID | / |
Family ID | 44062769 |
Filed Date | 2014-11-20 |
United States Patent
Application |
20140344048 |
Kind Code |
A1 |
Yonezaki; Tadashi ; et
al. |
November 20, 2014 |
METHODS AND APPARATUS FOR OPTIMIZING ADVERTISEMENT ALLOCATION
Abstract
In some embodiments, an apparatus includes a weight module, a
performance module and an allocator module. The weight module
calculates a weight for each segment from a set of segments of
potential advertisement placements matching a criterion. The weight
for a segment is based at least partially on (1) a budget score for
an advertisement campaign and (2) a number of potential placements
for the segment. The performance module calculates a performance
score for the advertisement campaign at each segment from the set
of segments. The performance score of the segment is based on a
success metric for an advertisement at the segment and a number of
impressions for the segment. The allocator module presents the
advertisement at a placement associated with the segment if the
weight for the segment is greater than a first threshold and the
performance score for the segment is greater than a second
threshold.
Inventors: |
Yonezaki; Tadashi; (Newton,
MA) ; Lee; Steven; (Arlington, MA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Tremor Video, Inc. |
New York |
NY |
US |
|
|
Assignee: |
TREMOR VIDEO, INC.
New York
NY
|
Family ID: |
44062769 |
Appl. No.: |
14/134606 |
Filed: |
December 19, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12950160 |
Nov 19, 2010 |
8615430 |
|
|
14134606 |
|
|
|
|
61281613 |
Nov 20, 2009 |
|
|
|
61384465 |
Sep 20, 2010 |
|
|
|
Current U.S.
Class: |
705/14.41 ;
705/14.48 |
Current CPC
Class: |
G06Q 30/0249 20130101;
G06Q 30/0242 20130101; G06Q 30/02 20130101 |
Class at
Publication: |
705/14.41 ;
705/14.48 |
International
Class: |
G06Q 30/02 20060101
G06Q030/02 |
Claims
1-20. (canceled)
21. A method, comprising: receiving a datum associated with an
advertisement campaign; calculating a budget score for the
advertisement campaign, the budget score being based on the datum
and a number of advertisement units in a campaign budget of the
advertisement campaign; calculating a weight for each segment from
a plurality of segments of potential placements matching a
criterion of the advertisement campaign, the weight for a segment
from the plurality of segments being based on the budget score and
a number of potential placements for that segment from the
plurality of segments; and sending a signal such that at least one
advertisement from the advertisement campaign is presented to a
placement associated with the segment from the plurality of
segments when the weight for that segment meets a criterion.
22. The method of claim 21, wherein the criterion is a first
criterion, the sending the signal includes sending the signal such
that the at least one advertisement from the advertisement campaign
is presented to the placement associated with the segment from the
plurality of segments when the weight for that segment meets the
first criterion and a predicted performance of the advertisement
campaign for that segment meets a second criterion.
23. The method of claim 21, further comprising: generating a random
value for the criterion.
24. The method of claim 21, further comprising: calculating a
predicted performance value for the segment from the plurality of
segments based on a relationship of a success metric for the
segment from the plurality of segments and a number of impressions
for the segment from the plurality of segments.
25. The method of claim 21, wherein the budget score for the
advertisement campaign increases as the number of advertisement
units in the campaign budget of the advertisement campaign
increases or the number of potential placements matching the
criterion of the advertisement campaign decreases.
26. The method of claim 21, wherein the sending the signal includes
sending the signal such that the at least one advertisement is
presented to a greater number of segments from the plurality of
segments as the budget score for the advertisement campaign
increases.
27. The method of claim 21, wherein the advertisement units include
at least one of advertisement impressions, advertisement clicks,
conversion rate, time spent, engagement rate, reach frequency, or
brand-lift.
28. The method of claim 21, further comprising: calculating the
number of advertisement units in the campaign budget, the number of
advertisement units being based on a total budget of the campaign
budget of the advertisement campaign and a cost of each
advertisement from the advertisement campaign.
29. An apparatus, comprising: a memory storing a weight module
configured to calculate a weight for each segment from a plurality
of segments of potential advertisement placements matching a first
criterion, the weight for a segment from the plurality of segments
being based at least partially on a number of potential placements
for the segment from the plurality of segments; the memory storing
a performance module configured to calculate a performance score
for the advertisement campaign at each segment from the plurality
of segments, the performance score of the segment from the
plurality of segments being based on a success metric for the at
least one advertisement at the segment from the plurality of
segments; and the memory storing an allocator module configured to
send a signal such that the at least one advertisement is presented
at a placement associated with the segment from the plurality of
segments when at least one of (1) the weight for the segment meets
a second criterion, or (2) the performance score for the segment
meets a third criterion.
30. The apparatus of claim 29, wherein: the memory stores a budget
module configured to calculate the budget score for the
advertisement campaign, the budget score being based on a number of
advertisement units in a campaign budget of the advertisement
campaign and a number of potential advertisement placements
matching the first criterion of the advertisement campaign.
31. The apparatus of claim 29, wherein: the memory stores a
placement estimator module configured to predict a number of
potential placements for each segment from the plurality of
segments.
32. The apparatus of claim 29, wherein the performance, module is
configured to assign a maximum performance score for the
performance score of the segment from the plurality of segments
prior to the at least one advertisement being placed at a placement
associated with the segment from the plurality of segments.
33. The apparatus of claim 29, wherein the allocator module is
configured to generate a random value for at least one of the
second criterion or the third criterion.
34. The apparatus of claim 29, wherein the weight for the segment
from the plurality of segments is at least partially based on a
ratio of the number of potential placements for the segment from
the plurality of segments to a number of potential placements for
the remaining segments from the plurality of segments.
35. The apparatus of claim 29, wherein the weight for the segment
from the plurality of segments is based at least partially on a
budget score for an advertisement campaign with at least one
advertisement.
36. The apparatus of claim 29, wherein the performance score of the
segment from the plurality of segments is based at least partially
on a number of impressions for the segment from the plurality of
segments.
37. A non-transitory processor-readable medium storing code
representing instructions to cause a processor to: receive a datum
associated with a campaign budget of an advertisement campaign;
calculate a weight for each segment from a plurality of segments of
potential advertisement placements matching a criterion, the weight
for a segment from the plurality of segments being based at least
partially on, the datum associated with the campaign budget;
calculate a predicted performance value for the segment from the
plurality of segments based on a success metric for the segment
from the plurality of segments; and send a signal to present an
advertisement from the advertisement campaign at a placement
associated with the segment from the plurality of segments based on
at least one of (1) the weight for the segment, or (2) the
predicted performance value for the segment.
38. The non-transitory processor-readable medium of claim 37,
wherein the code representing instructions to cause the processor
to send includes code representing instructions to cause the
processor to send the signal such that the advertisement from the
advertisement campaign is presented at the placement associated
with the segment from the plurality of segments when the weight for
the segment meets a first criterion and the predicted performance
value for the segment meets a second criterion.
39. The non-transitory processor-readable medium of claim 37,
wherein the success metric for the segment from the plurality of
segments includes a number of clicks for an advertisement from the
advertisement campaign displayed at a placement associated with the
segment from the plurality of segments.
40. The non-transitory processor-readable medium of claim 37,
wherein the campaign budget of the advertisement campaign provides
for at least one of a number of advertisement impressions or a
number of advertisement clicks.
41. The non-transitory processor-readable medium of claim 37,
wherein the predicted performance value for the segment from the
plurality of segments is a maximum predicted performance value
prior to an advertisement from the advertisement campaign being
placed at a placement associated with the segment from the
plurality of segments.
42. The non-transitory processor-readable medium of claim 37,
wherein the success metric for the segment from the plurality of
segments includes a number of items purchased and associated with
an advertisement from the advertisement campaign displayed at a
placement associated with the segment from the plurality of
segments.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. application Ser.
No. 12/950,160, filed Nov. 19, 2010, and entitled "Methods and
Apparatus for Optimizing Advertisement Allocation," which claims
priority to and the benefit of U.S. Provisional Patent Application
Ser. No. 61/281,613, filed Nov. 20, 2009, and entitled "Ad
Optimizer with Prediction Table," and U.S. Provisional Patent
Application Ser. No. 61/384,465, filed Sep. 20, 2010, and entitled
"Optimized Ad Allocation," each of which is incorporated herein by
reference in its entirety.
BACKGROUND
[0002] Embodiments described herein relate generally to
advertisement placement and more particularly to methods and
apparatus for optimizing advertisement allocation.
[0003] Some known advertisement allocators place advertisements at
placements (e.g., websites, video streams, audio streams, etc.)
based on a performance of the advertisement. For example, if the
advertisement has previously performed well, the advertisement
allocator will place the advertisement. Similarly, if the
advertisement has not previously performed well, the advertisement
allocator will not place the advertisement. Such known
advertisement allocators do not, however, account for a campaign
budget. As such, because of past performance of the advertisements
in a campaign, the advertisement allocator may not place enough
advertisements from the campaign to fill and/or use the budget
allotted to the advertisement campaign because of past
performance.
[0004] Accordingly, a need exists for methods and apparatus to
allocate advertisements from an advertisement campaign based on
performance and the budget allotted to the advertisement
campaign.
SUMMARY
[0005] In some embodiments, an apparatus includes a weight module,
a performance module and an allocator module. The weight module
calculates a weight for each segment from a set of segments of
potential advertisement placements matching a criterion. The weight
for a segment is based at least partially on (1) a budget score for
an advertisement campaign and (2) a number of potential placements
for the segment. The performance module calculates a performance
score for the advertisement campaign at each segment from the set
of segments. The performance score of the segment is based on a
success metric for an advertisement at the segment and a number of
impressions for the segment. The allocator module presents the
advertisement at a placement associated with the segment if the
weight for the segment is greater than a first threshold and the
performance score for the segment is greater than a second
threshold.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] FIG. 1 is a schematic diagram that illustrates communication
devices in communication with a host device via a network,
according to an embodiment.
[0007] FIG. 2 is a schematic illustration of a processor of a host
device, according to another embodiment.
[0008] FIG. 3 is a table illustrating an example of a campaign and
the segments associated with that campaign, according to another
embodiment.
[0009] FIG. 4 is a table illustrating an example of the entries of
a performance database for an advertisement campaign, according to
another embodiment.
[0010] FIG. 5 is a flow chart illustrating a method of optimizing
the allocation of advertisements, according to another
embodiment.
[0011] FIG. 6 is a flow chart illustrating another method of
optimizing the allocation of advertisements, according to another
embodiment.
DETAILED DESCRIPTION
[0012] In some embodiments, an apparatus includes a weight module,
a performance module and an allocator module. The weight module is
configured to calculate a weight for each segment from a set of
segments of potential advertisement placements matching a
criterion. The weight for a segment from the set of segments is
based at least partially on (1) a budget score for an advertisement
campaign with at least one advertisement and (2) a number of
potential placements for the segment from the set of segments. The
performance module is configured to calculate a performance score
for the advertisement campaign at each segment from the set of
segments. The performance score of the segment from the set of
segments is based on a success metric for the at least one
advertisement at the segment from the set of segments and a number
of impressions for the segment from the set of segments. The
allocator module is configured to present the at least one
advertisement at a placement associated with the segment if the
weight for the segment is greater than a first threshold and the
performance score for the segment is greater than a second
threshold.
[0013] In such embodiments, the apparatus can ensure that the
budget of an advertising campaign is spent by using a budget score
when calculating the weight of a segment. Additionally, the
apparatus can ensure that advertisements are not placed at segments
having low performance unless too few segments having a high
performance are available for placement such that the campaign
budget is spent. Accordingly, the apparatus places advertisements
at lower performing segments when the budget would not be spent by
placing advertisements solely at higher performing segments.
[0014] A non-transitory processor-readable medium stores code that
represents instructions to cause a processor to calculate a budget
score for an advertisement campaign. The budget score is based on a
number of advertisement units in a campaign budget of the
advertisement campaign and a number of potential placements
matching a criterion of the advertisement campaign. The
non-transitory processor-readable medium stores code that
represents instructions to cause the processor to calculate a
weight for each segment from a set of segments of the potential
placements matching the criterion. The weight for a segment from
the set of segments is based on the budget score and a relationship
between a number of potential placements for that segment from the
set of segments and a number of potential placements for the
remaining segments from the set of segments. The non-transitory
processor-readable medium further stores code that represents
instructions to cause the processor to present at least one
advertisement from the advertisement campaign to a placement
associated with the segment from the set of segments if the weight
for that segment is greater than a threshold.
[0015] A non-transitory processor-readable medium stores code that
represents instructions to cause a processor to calculate a weight
for each segment from a set of segments of potential advertisement
placements matching a criterion. The weight for a segment from the
set of segments is based at least partially on a campaign budget of
an advertisement campaign and a number of potential placements for
the segment from the set of segments. The non-transitory
processor-readable medium stores code that represents instructions
to cause the processor to calculate a predicted performance value
for the segment from the set of segments based on a relationship of
a success metric for the segment from the set of segments and a
number of impressions for the segment from the set of segments. The
non-transitory processor-readable medium further stores code that
represents instructions to cause the processor to determine whether
to present an advertisement from the advertisement campaign at a
placement associated with the segment from the set of segments
based on the weight for the segment and the predicted performance
value for the segment.
[0016] As used herein, "criterion" can include a criterion defined
by a single property, parameter and/or requirement and a criterion
defined by multiple properties, parameters and/or requirements. For
example, a first criterion can include "all males" and a second
criterion can include "all females between the ages of 55-65 living
in New York City."
[0017] As used in this specification, the singular forms "a," "an"
and "the" include plural referents unless the context clearly
dictates otherwise. Thus, for example, the term "module" is
intended to mean a single module or a combination modules.
[0018] FIG. 1 is a schematic diagram that illustrates communication
devices 180 in communication with a host device 120 via a network
170, according to an embodiment. Specifically, communication device
150 is configured to communicate with the host device 120.
Similarly, communication device 160 is configured to communicate
with the host device 120. The network 170 can be any type of
network (e.g., a local area network (LAN), a wide area network
(WAN), a virtual network, a telecommunications network) implemented
as a wired network and/or wireless network. As described in further
detail herein, in some embodiments, for example, the communication
devices 180 are personal computers connected to the host device 120
via an internet service provider (ISP) and the Internet (e.g.,
network 170).
[0019] The host device 120 can be any type of device configured to
send data over the network 170 to and/or receive data from one or
more of the communication devices 180. In some embodiments, the
host device 120 can be configured to function as, for example, a
server device (e.g., a web server device), a network management
device, an advertisement placement device and/or so forth.
[0020] The host device 120 includes a memory 124 and a processor
122. The memory 124 can be, for example, a random access memory
(RAM), a memory buffer, a hard drive, a database, an erasable
programmable read-only memory (EPROM), an electrically erasable
read-only memory (EEPROM), a read-only memory (ROM) and/or so
forth. In some embodiments, the memory 124 of the host device 120
includes data used to place advertisements at various locations
(e.g., at various websites). In such embodiments, for example, the
host device 120 is configured to place advertisements within video
streams, within audio streams, as pop-up advertisements, as banner
advertisements, as advertisements embedded in the text of a website
and/or the like. In some embodiments, the memory 124 stores
instructions to cause the processor to execute modules, processes
and/or functions.
[0021] The processor 122 of the host device 120 can be any suitable
processing device configured to perform advertisement optimization
and place advertisements at optimal placements (e.g., websites,
video streams, audio streams, etc.), as described in further detail
herein. More specifically, as described in further detail herein,
the processor 122 can be configured to execute modules, functions
and/or processes to optimize the placement of advertisements. In
some embodiments, the processor 122 can be a general purpose
processor, a Field Programmable Gate Array (FPGA), an Application
Specific Integrated Circuit (ASIC), a Digital Signal Processor
(DSP), and/or the like.
[0022] The host device 120 is operatively coupled to a placement
database 172 and a performance database 174. The placement database
172 and the performance database 174 can be any suitable databases
such as, for example, relational databases, object databases,
object-relational databases, hierarchical databases, network
databases, entity-relationship databases, and/or the like. While
shown in FIG. 1 as being separate from the host device 120, in
other embodiments the placement database 172 and the performance
database 174 can be part of the host device 120. For example, the
placement database 172 and the performance database 174 can be
stored in the memory 124. Additionally, while shown in FIG. 1 as
being separate databases, in other embodiments the placement
database 172 and the performance database 174 can be part of a
single database. As described in further detail herein, the host
device 120 can use the placement database 172 and the performance
database 174 to optimize the placement of the advertisements while
using a budget of an advertisement campaign.
[0023] The placement database 172 is configured to store and/or
maintain data associated with possible advertisement placements.
More specifically, the placement database 172 stores and/or
maintains data associated with each web site, video stream and/or
audio stream at which the host device 120 can place and/or embed an
advertisement. In some embodiments, the placement database 172 can
also store and/or maintain data associated with probable
demographics of each placement. For example, if a placement is a
cartoon video stream, the demographic associated with that
placement might be children under thirteen. For another example, if
the placement is a football website, the demographic associated
with that placement might be males over twenty. In some
embodiments, and as described in further detail herein, a placement
can be part of and/or associated with multiple demographic
categories. Additionally, as described in further detail herein, a
placement can be associated with sub-categories and/or demographic
segments. For example, a placement can be associated with both
"individuals between the age of 25-45" and "dog owners in
Washington, D.C. between the age of 30-40." In such an example, the
segment "dog owners in Washington, D.C. between the age of 30-40"
is a sub-category and/or demographic segment of the broader
demographic category "individuals between the age of 25-45."
[0024] The performance database 174 is configured to store and/or
maintain data associated with the performance of specific
advertisements at specific placement categories and/or segments.
Such performance can be calculated and/or determined using any
suitable metric. In some embodiments, for example, the performance
of an advertisement at a segment can be a clickthrough rate (CTR)
of the advertisement at the segment, a conversion rate of the
advertisement at the segment, an engagement rate of the
advertisement at the segment, a reach of the advertisement at the
segment, the brand-lift of the advertisement at the segment, and/or
the like. As described in further detail herein, the performance
data stored in the performance database 174 can be used to predict
future performance of an advertisement at a placement associated
with a segment.
[0025] Each of the communication devices 180 can be, for example, a
computing entity (e.g., a personal computing device such as a
desktop computer, a laptop computer, etc.), a mobile phone, a
monitoring device, a personal digital assistant (PDA), and/or so
forth. Although not shown, in some embodiments, each of the
communication devices 180 can include one or more network interface
devices (e.g., a network interface card) configured to connect the
communication devices 180 to the network 170. In some embodiments,
the communication devices 180 can be referred to as client
devices.
[0026] As shown in FIG. 1, the communication device 160 has a
processor 162, a memory 164, and a display 166. The memory 164 can
be, for example, a random access memory (RAM), a memory buffer, a
hard drive, and/or so forth. The display 166 can be any suitable
display, such as, for example, a liquid crystal display (LCD), a
cathode ray tube display (CRT) or the like. Similar to
communication device 160, the communication device 150 has a
processor 152, a memory 154, and a display 156.
[0027] In some embodiments, a web browser application can be stored
in the memory 164 of the communication device 160. Using the web
browser application, the communication device 160 can send data to
and receive data from the host device 120. Similarly, the
communication device 150 can include a web browser application. In
such embodiments, the communication devices 180 act as thin
clients. This allows minimal data to be stored on the communication
devices 180. In other embodiments, the communication devices 180
can include an application specific to communicating with the host
device 120.
[0028] In some embodiments, when a user of a communication device
180 accesses a website using the web browser application, the host
device can determine which advertisement to present to the user. As
described in further detail herein, such a determination can be
based on the budget of an advertisement campaign, the predicted
performance of an advertisement at one or more segments with which
the website is associated, a target demographic and/or criterion
specified by the advertisement campaign, a number of potential
placements for the one or more segments, a number of potential
placements for the target demographic and/or criterion specified by
the advertisement campaign, and/or the like.
[0029] As discussed above, the communication devices 180 can send
data to and receive data from the host device 120 associated with
advertisements. In some embodiments, the data sent between the
communication devices 180 and the host device 120 can be formatted
using any suitable format. In some embodiments, for example, the
data can be formatted using extensible markup language (XML),
hypertext markup language (HTML) and/or the like.
[0030] In some embodiments, one or more portions of the host device
120 and/or one or more portions of the communication devices 180
can include a hardware-based module (e.g., a digital signal
processor (DSP), a field programmable gate array (FPGA)) and/or a
software-based module (e.g., a module of computer code to be
executed at a processor, a set of processor-readable instructions
that can be executed at a processor). In some embodiments, one or
more of the functions associated with the host device 120 (e.g.,
the functions associated with the processor 122) can be included in
one or more modules (see, e.g., FIG. 2). In some embodiments, one
or more of the functions associated with the communication devices
180 (e.g., functions associated with processor 152 or processor
162) can be included in one or more modules. In some embodiments,
one or more of the communication devices 180 can be configured to
perform one or more functions associated with the host device 120,
and vice versa.
[0031] In some embodiments, the host device 120 does not deliver
and/or provide advertisements directly to the communication devices
180. In such embodiments, for example, the host device 120 provides
the advertisements to a third party (e.g., the owner of a website
including a potential advertisement placement). The third party can
then deliver and/or provide the advertisement to the communication
devices 180 with the other content of the website. In other
embodiments, the host device 120 delivers and/or provides the
advertisements directly to the communication devices 180. In such
embodiments, the third party can provide a website and/or other
content to the communication devices 180 such that the
advertisement can be presented to the user with the content
provided by the third party.
[0032] FIG. 2 is a schematic illustration of a processor 200 of a
host device, according to another embodiment. In some embodiments,
the processor 200 can be similar to the processor 122 of the host
device 120. More specifically, the processor 200 can be any
suitable processing device configured to perform advertisement
optimization and allocate advertisements to be placed at optimal
placements.
[0033] The processor 200 includes a budget module 202, a placement
estimator module 204, a weight module 206, a performance module
208, and an allocator module 210. While not shown in FIG. 2, in
some embodiments, the processor 200 can include a communication
module configured to communicate with communication devices (e.g.,
communication devices 180), third party host devices (e.g., website
host devices), and/or other modules and/or devices.
[0034] The budget module 202 can be configured to calculate and/or
determine a budget score for an advertisement campaign. In some
embodiments, the budget score can be based on a number of
advertisement units in a campaign budget of an advertisement
campaign (e.g., N.sub.Budget). In some embodiments, the
advertisement units can be a number of impressions, a number of
advertisement clicks, a conversion rate, a time spent, an
engagement rate, a specified reach, a specified frequency, a
specified brand-lift measurement, and/or the like, included within
the campaign budget. For example, if an advertisement costs $1.00
per advertisement impression and a campaign budget is $1000, the
advertisement campaign can include 1000 impressions.
[0035] In some embodiments, the budget score is also based on a
number of potential placements (e.g., websites, video streams,
audio streams, etc.) matching a criterion of the advertisement
campaign (e.g., N.sub.Campaign). In such embodiments, a provider of
the advertisement campaign can specify a target demographic and/or
criterion for the advertisement campaign. For example, a company
might want to target individuals between the ages 25-45. For
another example, a sports car company might want to initiate an
advertisement campaign directed toward males between the ages of
35-55 during evening hours. Such a target can be supplied by the
provider and/or purchaser of the advertisement campaign.
[0036] The placement estimator module 204 can be configured to
estimate and/or determine a number of placements that are
associated with and/or correspond to a supplied demographic and/or
criterion. For example, the placement estimator module can receive
the target demographic and/or criterion supplied by the provider
and/or purchaser of the advertisement campaign and determine the
number of potential placements that are associated with and/or
correspond to that target demographic and/or criterion. Using the
example provided above, the placement estimator module 204 can
estimate and/or determine that there are 100 possible placements
for the demographic/criterion: "individuals between the ages
25-45."
[0037] In some embodiments, the budget score for an advertisement
campaign can be computed using formula (1):
Budget Score = { 1 if N Campaign .ltoreq. 0 or N Budget < 0 N
Budget N Campaign otherwise ( 1 ) ##EQU00001##
As discussed above, N.sub.Budget can be a number of advertisement
units in a campaign budget of an advertisement campaign and
N.sub.Campaign can be a number of potential placements matching a
criterion of the advertisement campaign. For example, if
N.sub.Budget=1000 (e.g., 1000 impressions in the budget) and
N.sub.Campaign=100 (e.g., 100 possible placements for the
criterion), the budget score is equal to 10. Accordingly, the
budget score is a relationship between a number of advertisement
units in a campaign budget (N.sub.Budget) and a number of potential
placements matching a criterion of the advertisement campaign
(N.sub.Campaign). For another example, if there are no possible
placements for a given criterion (i.e., N.sub.Campaign=0) or if the
advertisement campaign is over budget (i.e., N.sub.Budget<0),
the budget score equals 1. In other embodiments, any other suitable
relationship between the number of advertisement units in a
campaign budget (N.sub.Budget) and the number of potential
placements matching a criterion of the advertisement campaign
(N.sub.Campaign) can be used to calculate a budget score.
[0038] The weight module 206 can be configured to calculate a
weight for the campaign at each segment of the advertisement
campaign. FIG. 3, for example, is a table 300 illustrating the
segments of an advertisement campaign. The criterion of the
advertisement campaign provided by the purchaser and/or provider of
the advertisement campaign is "individuals between the age 25-45."
As illustrated in the table 300, the number of potential placements
for the entire campaign is 100. These potential placements can be
divided into multiple segments. In this example, the potential
placements are divided into four unique segments. Segment 1, for
example, includes placements (e.g., websites, video streams, audio
streams, etc.) targeting "males having an income of greater than
$100,000." As shown in FIG. 3, Segment 1 includes 20 possible
placements. Similarly, Segment 2 includes placements targeting
"females located in New York City" and includes 10 possible
placements; Segment 3 includes placements targeting "males having
an income of less than or equal to $100,000" and includes 40
possible placements; and Segment 4 includes placements targeting
"females living outside New Your City" and includes 30 possible
placements. Each segment of the advertisement campaign is a subset
of the campaign criterion ("individuals between the age
25-45").
[0039] While FIG. 3 includes an example of a campaign including
four segments, in other embodiments, a campaign can include any
number of segments based on any suitable criterion, such as, for
example, age, sex, geographic location, income level, indicated
preferences, websites visited, internet protocol (IP) address,
education level, school attended, website subscriptions,
profession, interests, hobbies, time of the day, day of the week,
month, season of year, and/or the like. Additionally, while shown
in FIG. 3 as being mutually exclusive of each other, in other
embodiments, a predicted placement can be included in more than one
segment. As such, the criterion of one segment can overlap the
criterion of another segment and/or a placement can be applicable
to and/or associated with multiple segments.
[0040] Returning to FIG. 2, the weight module 206 can use the
segment information and the budget score (calculated by the budget
module 202) to calculate a weight for each segment. In some
embodiments, the weight module 206 can use the following formulas
(formulas (2) and (3)) to calculate the weight for each
segment:
Segment Weight = { 1 if N Other = 0 and Budget_Score .gtoreq. 1
Budget Score .times. N Segment N Other if N Other > 0 Budget
Score .times. N Segment otherwise ( 2 ) N Other = N Campaign - N
Segment ( 3 ) ##EQU00002##
[0041] As discussed above, N.sub.Campaign can be a number of
potential placements matching a criterion of an advertisement
campaign and the budget score can be calculated by the budget
module 202. N.sub.Segment can be a number of potential placements
matching a criterion of a segment of the advertisement campaign.
For example, N.sub.Segment for Segment 1 of the table 300 of FIG. 3
equals 20, N.sub.Segment for Segment 2 equals 10, N.sub.Segment for
Segment 3 equals 40, and N.sub.Segment for Segment 4 equals 30.
[0042] Accordingly, if N.sub.Other equals zero and the budget score
is greater than or equal to 1, the segment weight for that segment
will be 1. According to formula (2), if N.sub.Other equals 0, every
potential placement in an advertisement campaign is included within
a single segment (i.e., N.sub.Campaign=N.sub.Segment). Thus,
depending on the value of the budget score, the segment weight for
a segment when N.sub.Other is 0 will equal 1 if the budget score is
greater than or equal to 1 or the budget score multiplied by the
number of placements in the segment (N.sub.Segment) if the budget
score is less than 1. In some embodiments, a budget score greater
than one indicates that the number of advertising units in a
campaign budget is greater than or equal to the number of possible
placements for that budget.
[0043] Using the example of FIG. 3, a segment weight (i.e., a raw
segment weight) can be calculated for each of the segments. For
example, N.sub.Other of Segment 1 equals 80 (100 total placements
for the campaign--20 total placements for Segment 1). N.sub.Other
is greater than 0 and, thus, the segment weight for the Segment 1
is based on the budget score for the campaign as well as a
relationship between the number of placements for Segment 1 and the
number of placements for the other segments of the campaign (i.e.,
Segment 2, Segment 3 and Segment 4). More specifically, using the
budget score of 10 (e.g., 1000 possible impressions in the budget
and 100 possible placements for the entire campaign) the segment
weight for Segment 1 is equal to 2.5 (i.e., 10.times.20/80). Using
a similar calculation, the segment weight (i.e., raw segment
weight) for Segment 2 is equal to 1.11 (i.e., 10.times.10/90), the
segment weight for Segment 3 is equal to 6.67 (i.e.,
10.times.40/60), and the segment weight for Segment 4 is equal to
4.29 (i.e., 10.times.30/70).
[0044] In some embodiments, the raw weights for the segments can be
normalized with respect to the sum of the raw weights for all the
segments of a campaign. In such embodiments, for example, the
normalized weight for Segment 1 can be calculated by dividing the
raw weight of Segment 1 (2.5) by the sum of the raw segment weights
for the segments of the campaign (i.e., 2.5+1.11+6.67+4.29=14.57).
Accordingly, the normalized weight for Segment 1 equals 0.172. The
normalized weight for Segment 2 (i.e., 1.11/14.57=0.076), the
normalized weight for Segment 3 (i.e., 6.67/14.57=0.458), and the
normalized weight for Segment 4 (i.e., 4.29/14.57=0.294) can be
similarly calculated.
[0045] After the weights are normalized, the weight module 206 can
provide to the allocator module 210 the normalized weight for each
segment. In other embodiments, the raw weights are not normalized
and the weight module 206 provides the raw weights to the allocator
module 210.
[0046] The performance module 208 can be configured to calculate a
performance score for each segment from the segments of the
advertising campaign. The performance score of a segment can be
used as an indication of a probable performance (i.e., predicted
performance) of the advertisements of an advertisement campaign
presented at the placements of that segment. The performance score
of each segment can be calculated using the following formula
(formula (4)):
Performance Score = Success Metric + 1 Number of Impressions + 1 (
4 ) ##EQU00003##
The success metric for a segment can be a number of clicks, a
number of items purchased based on the advertisement, and/or the
like. The number of impressions for a segment is the number of
advertisement impressions (e.g., views, audio playbacks, video
playbacks, etc.) for an advertising campaign placed at the segment.
Accordingly, the performance score is a modified ratio of the
success metric and the number of impressions.
[0047] In formula (4), one is added to both the success metric and
the number of impressions to ensure that a performance score for a
segment is not initially zero. According to formula (4), the
performance score for a segment is initially one (i.e., the maximum
value for the performance score). As described in further detail
herein with respect to the allocator module 210, this ensures that
advertisements associated with an advertisement campaign are
presented at a new advertisement segment. Similarly stated, a
segment initially is provided a high performance score rather than
a low performance score. As the number of impressions increases
without the success metric increasing, the performance score begins
to decrease. Thus, after the number of impressions becomes
significantly large (i.e., the sample size increases), the
performance score of an unsuccessful segment (i.e., as measured by
the success metric) decreases. Similarly, as the success metric
increases (e.g., the number of clicks for an advertisement at a
segment increases), the performance score of that segment will
remain high and/or will increase.
[0048] In some embodiments, the performance module 208 can retrieve
the success metric and the number of impressions for a segment from
a performance database (e.g., performance database 174 of FIG. 1).
In such embodiments, the performance database can store a success
metric and a number of impressions for each segment associated with
an advertisement campaign. Continuing with the above referenced
example, FIG. 4 is a table illustrating an example of the entries
of a performance database for the advertisement campaign. Based on
the values associated with each segment, the performance module 208
can calculate a performance score for each segment. For example,
the performance score for Segment 1 is 0.505 (i.e.,
(50+1)/(100+1)), the performance score for Segment 2 is 0.189
(i.e., (20+1)/(110+1)), the performance score for Segment 3 is
0.333 (i.e., (0+1)/(2+1)), and the performance score for Segment 4
is 0.294 (i.e., (14+1)/(50+1)).
[0049] In some embodiments, the performance database can be updated
each time the success metric and/or the number of impressions for a
segment increases. Accordingly, the performance score for each
segment can be current. In other embodiments, the performance
database is updated periodically (e.g., every 10 seconds), after a
predetermined number of impressions (e.g., after every 10
impressions for the campaign), and/or the like.
[0050] In some embodiments, after a period of time has elapsed,
data (e.g., success metrics and/or number of impressions) can be
removed from the performance database 174. For example, all data
that was collected more than a time period (e.g., one week, one
month, one year, etc.) before the current time can be removed from
the performance database 174. This can ensure that the performance
score for each segment is calculated based on current (and not
outdated) data.
[0051] In some embodiments, the results of the calculations
performed by the performance module 208 can be stored in a look-up
table. In some embodiments, such a look-up table can be stored at a
memory collocated with the processor 200 (e.g., memory 124
collocated with processor 122 in FIG. 1). In other embodiments,
such a look-up table can be stored at a database such as the
performance database 174. Such a look-up table can be used by the
processor 200 to quickly retrieve a performance score for a segment
(i.e., without having to calculate the performance score each time
the campaign determines whether to place an advertisement at a
segment).
[0052] In some embodiments, the look-up table can use a compression
scheme and/or method to compress the size of the look-up table.
This allows the look-up table (which can become relatively large)
to be stored using less memory. In some embodiments, for example, a
modified Run-Length-Encoding (RLE) scheme can be used to compress
the look-up table. In such embodiments, an RLE compression scheme
can be modified such that each coefficient of the RLE values is a
cumulative coefficient rather than a discrete coefficient. For
example, for a data value of "AAAABBAAAAAABBBA," RLE compression
would result in "4A2B6A3B1A." In a modified RLE scheme, the same
data value would result in "4A6B12A15B16A." Such a modified RLE
scheme results in less computation as the coefficients are
cumulative. Thus, for example, to determine the 13.sup.th bit in
the data value, a processor looks to which coefficients 13 is
between. In this example, 13 is between the coefficients 12 and 15.
Accordingly, the value of the 13.sup.th bit is "B," the value
following the higher coefficient (i.e., 15). In other embodiments,
any other suitable compression scheme and/or method can be used to
compress the look-up table.
[0053] After the performance module 208 calculates the performance
score for each segment, the performance module 208 can send the
performance scores to the allocator module 210. Using the
performance scores for each segment and the normalized (or raw)
weight for each segment, the allocator module 210 can determine
whether or not to provide an advertisement associated with an
advertisement campaign to a placement associated with a the
segments.
[0054] In some embodiments, for example, the allocator module 210
can compare the weight of each segment with a weight threshold. If
the weight of a segment is less than the weight threshold, for
example, the allocator module 210 does not place an advertisement
at that segment. If the weight of a segment is greater than the
weight threshold, the allocator can further consider whether to
place an advertisement at the segment. Because the allocator module
uses the weight of a segment to at least partially determine
whether or not to provide an advertisement to a placement of a
segment, there is a greater probability that advertisements will be
provided to the segments with the greater number of possible
placements. This helps to ensure that the budget of the
advertisement campaign is spent by providing advertisements to the
segments with the greater number of possible placements. In some
embodiments, the normalized weights are used to calculate the
weight threshold and/or are compared to the weight threshold. In
other embodiments, the raw weights are used to calculate the weight
threshold and/or are compared to the weight threshold.
[0055] In some embodiments, the weight threshold can be
predetermined. For example, if the normalized weight of a segment
is less than 0.20, an advertisement is not placed at the segment.
Continuing the example described above, in such an embodiment, the
allocator module 210 does not place advertisements at Segment 1
(normalized weight=0.172) or Segment 2 (normalized weight=0.076)
because their normalized weight is below the weight threshold
(0.20).
[0056] In other embodiments, the weight threshold can be a random
value between the lowest normalized weight and the highest
normalized weight of the segments of an advertisement campaign. For
the above described example, the weight threshold can be a random
value between 0.076 (the normalized weight for Segment 2) and 0.458
(the normalized weight for Segment 3). In still other embodiments,
the weight threshold value can be an average of the raw weights.
For example, the average raw weight for Segment 1, Segment 2,
Segment 3 and Segment 4 is 3.64 (i.e., (2.5+1.11+6.67+4.29)/4). In
such embodiments, the raw weights of the segments can be compared
to the weight threshold.
[0057] If the weight of the segment is greater than the weight
threshold, the allocator module 210 can compare the performance
score of the segment to a performance threshold. If the performance
score of the segment is greater than the threshold, the segment can
be placed in a pool and/or group of segments at which
advertisements from an advertisement campaign will be placed.
[0058] In some embodiments, the performance threshold can be
predetermined. In such embodiments, an advertisement campaign
provider can determine at what level of predicted performance
(i.e., using the performance scores) to provide an advertisement to
a segment. In other embodiments, the performance threshold can be
determined based on the performance of the segments of the
advertisement campaign. In some embodiments, for example, the
performance threshold can be an average performance score of the
segments from the advertisement campaign. In the above described
example, the average performance score of the segments is 0.330
(i.e., (0.505+0.189+0.333+0.294)/4). Accordingly, in this example,
the performance score of Segment 1 and the performance score of
Segment 3 are above the performance threshold. In other
embodiments, the performance threshold can be any other suitable
value such as, for example, a ratio of the total success of the
advertisement campaign and the total number of impressions for the
advertisement campaign, a random value, a median performance score
of the segments of the advertisement campaign, and/or the like.
[0059] The allocator module 210 can then place advertisements at
the placements associated with the segments having both a weight
greater than the weight threshold and a performance score greater
than the performance threshold. The allocator module 210 can place
the advertisements with the segments having the highest performance
score first, and those having the lowest performance score last.
Thus, if the campaign budget runs out prior to advertisements being
placed at each segment having both a weight greater than the weight
threshold and a performance score greater than the performance
threshold, advertisements will not be placed at placements with the
lowest performance score. Similarly stated, advertisements are
placed at placements with the lowest performance score if a
sufficient number of placements with higher performance scores are
not available (e.g., advertisements have been placed at the
placements with higher performance scores but budget sill
remains).
[0060] After placement, the performance score for each segment can
be continually updated. Similarly, the budget score for each
campaign and the weights for each segment can be continually
updated as more placements become available for a campaign and/or a
segment and as the budget of a campaign increases and/or decreases.
Accordingly, the host device can use the budget of the
advertisement campaign in an effective manner by using segments
having enough possible placements to ensure that the budget is
spent but by placing the segments at the highest performing
placements.
[0061] Because the weight of each segment is based partially on the
budget score of the advertisement campaign, as the budget increases
for an advertisement campaign and/or the number of possible
placements for a campaign decreases, the deviation of the weights
of the segments (i.e., difference between the highest segment
weight and the lowest segment weight) increases. As described in
further detail herein, this ensures that an advertisement campaign
with a large budget but with few possible placements is spent on
and/or focused at segments with the largest number of possible
placements. Similarly, as the budget decreases for an advertisement
campaign and/or the number of possible placements for a campaign
increases, the deviation of the weights of the segments decreases.
This increases the probability that the placement of advertisements
at segments will be based more on predicted performance (i.e.,
performance scores) than on volume (i.e., which segment has the
greatest number of possible placements).
[0062] While shown and described above as being based solely on a
single campaign, in other embodiments, the weight for each segment
associated with a campaign can be calculated with and/or depend at
least in part on the weight for that segment associated with
another campaign. For example, Segment 1 can be applicable to any
number of campaigns. For example, Segment 1 can be applicable to a
campaign with the criterion of "Age: over 50", to a campaign with
the criterion of "Location: Boston", or any other campaign having a
criterion of which "Male; income >$100k" can be a subset.
Accordingly, each of these campaigns will have advertisements that
can be placed at placements associated with Segment 1. Making a
weight for a segment associated with a campaign depend on the other
campaigns with which that segment is associated ensures that the
advertisements from optimal campaigns will be placed at the
placements associated with that segment.
[0063] In some embodiments the following formulas (5) and (6) can
be used (instead of the formulas (2) and (3) described above) to
calculate the weight for each segment with respect to a
campaign:
if N Other ( AC ) = 0 and Budget_Score ( AC ) .gtoreq. 1 Segment
Weight ( CC ) = { 0 if N Other ( CC ) > 0 1 otherwise else
Segment Weight ( CC ) = { Budget Score ( CC ) .times. N Segment N
Other ( CC ) if N Other ( CC ) > 0 Budget Score ( CC ) .times. N
Segment otherwise ( 5 ) N Other ( X ) = N Campaign ( X ) - N
Segment ( 6 ) ##EQU00004##
For formulas (5) and (6):
AC=Any Campaign
CC=Current Campaign
[0064] As discussed above, N.sub.Campaign can be a number of
potential placements matching a criterion of an advertisement
campaign and the budget score for each campaign can be calculated
by the budget module 202. N.sub.Segment can be a number of
potential placements matching a criterion of a segment. For
example, N.sub.Segment for Segment 1 of the table 300 of FIG. 3
equals 20, N.sub.Segment for Segment 2 equals 10, N.sub.Segment for
Segment 3 equals 40, and N.sub.Segment for Segment 4 equals 30.
[0065] Using formula (5), if N.sub.Other(AC) (i.e., N.sub.Other for
any campaign associated with a common segment) equals zero and the
budget score for that campaign (Budget_Score.sub.(AC) is greater
than or equal to 1, the segment weight for that segment associated
with a current campaign (i.e., Segment Weight.sub.(CC) will be 0 if
N.sub.Other(CC) for the current campaign is greater than 0.
Otherwise the segment weight for that segment associated with the
current campaign (i.e., Segment Weight.sub.(CC) equals 1.
Accordingly, the Segment Weight for a segment of a campaign, in
which that segment is the only segment in that campaign (i.e.,
N.sub.Campaign=N.sub.Segment) and the budget score for that
campaign is greater than or equal to 1 (as calculated above), will
be equal to 1 while the segment weights for that segment of other
campaigns (i.e., campaigns having more than the one segment) will
be equal to 0. As described in further detail herein, this ensures
that advertisements associated with that campaign (i.e., the
campaign having N.sub.Campaign=N.sub.Segment) will be placed at the
placements associated with that segment while advertisements
associated with the other campaigns (i.e., campaigns having more
than the one segment--N.sub.Campaign>N.sub.Segment) will not be
placed at the placements associated with that segment. Similarly
stated, because that segment is the only segment associated with
that campaign, the advertisements associated with that campaign are
given priority over advertisements associated with other campaigns,
to be placed at placements associated with that segment.
[0066] If N.sub.Other(AC) (i.e., N.sub.Other for any campaign
associated with a segment) does not equal zero or if the budget
score for a campaign having N.sub.Other(AC) equal to zero
(Budget_Score.sub.(AC)) is greater than or equal to 1, the segment
weight for that segment associated with a current campaign (i.e.,
Segment Weight.sub.(CC) can be calculated similar to calculating
the segment weights using formula (2).
[0067] While shown and described above as normalizing raw weights
with respect to the segments in a single campaign, in other
embodiments, the raw weights can be normalized using any other
suitable method. For example, in some embodiments, the raw weights
for the segments can be normalized with respect to the sum of the
raw weights for that segment across multiple campaigns. More
specifically, as discussed above, Segment 1 can be applicable to
any number of campaigns. For example, if Segment 1 (FIG. 3) is
associated with two campaigns (Campaign 1 and Campaign 2), the raw
weights of Segment 1 with respect to the two campaigns can be
normalized. For example, if the raw weight of Segment 1, Campaign 1
is 2.5 (calculated above with respect to formulas (2) and (3)) and
the raw weight of Segment 1, Campaign 2 is 5, the normalized weight
for Segment 1, Campaign 1 is 0.333 (2.5/(2.5+5)). Similarly, the
normalized weight for Segment 1, Campaign 2 is 0.666 (5/2.5+5)).
Accordingly, as discussed in further detail herein, because Segment
1, Campaign 2 has a greater weight than Segment 1, Campaign 1, a
greater number of advertisements from Campaign 2 will be placed at
the placements of Segment 1 than advertisements from Campaign 1.
Thus, the normalized weights ensure that a specific segment (e.g.,
Segment 1) is used efficiently among the different campaigns (e.g.,
Campaign 1 and Campaign 2). More specifically, the normalized
weights ensure that advertisements placed at the placements of a
segment are from campaigns with higher weights for that
segment.
[0068] FIG. 5 is a flow chart illustrating a method 400 of
optimizing the allocation of advertisements. In some embodiments,
the method 400 can be performed by a processor (e.g., processor 122
or 200) at a host device (e.g., host device 120). As such, the
processor can provide, using the method 400, an indication as to
what advertisements to present, place and/or serve with respect to
an advertisement campaign.
[0069] The method 400 includes calculating a number of
advertisement units in a budget, at 402. As described above, in
some embodiments advertisement units included within the campaign
budget can be a number of impressions, a number of advertisement
clicks, a conversion rate, a time spent, an engagement rate, a
specified reach, a specified frequency, a specified brand-lift
measurement, and/or the like. Accordingly, a total budget amount
can be divided by a cost per advertisement unit to determine the
number of advertisement units in the budget.
[0070] A number of possible placements for a campaign is
calculated, at 404. In some embodiments and as described above, the
number of possible placements for a campaign can be calculated by a
placement estimator module (e.g., placement estimator module 204,
shown and described with respect to FIG. 2). In some embodiments, a
campaign can be targeted toward a specific demographic and/or
criterion (e.g., age, sex, geographic location, income level,
indicated preferences, websites visited, IP address, education
level, school attended, website subscriptions, profession,
interests, hobbies, time of the day, day of the week, month, season
of year, and/or the like). The number of possible placements for a
campaign is the number of possible places (e.g., websites, audio
streams, video streams, etc.) to present an advertisement according
to the targeted demographic and/or criterion. Similarly stated, the
number of possible placements for a campaign is the amount of
placement inventory for the specified criterion of a campaign.
[0071] A campaign score is calculated, at 406. The campaign score
is calculated based on the number of advertisement units in a
budget and the number of possible placements for a campaign. In
some embodiments, the campaign score can be a relationship (e.g., a
ratio) between the number of advertisement units in the budget and
the number of possible placements for the campaign. Thus, the
campaign score can provide an indication of the number of
advertisement units for the budget with respect to the number of
possible placements at which the advertisement units can be
obtained (e.g., clicks, impressions, etc.). In some embodiments,
and as described above, such a calculation can be performed by a
budget module (e.g., budget module 202, shown and described with
respect to FIG. 2).
[0072] A number of possible placements for each segment associated
with the campaign can be calculated, at 408. The criterion of a
segment of a campaign can be a subset of the criterion of the
campaign. For example, if the campaign is directed toward
individuals with an income of greater than $100,000, a segment of
that campaign could be females between 35-45 living in the pacific
time zone with an income of greater than $100,000. A criterion for
a segment can have any number of properties, parameters, and/or
requirements. The number of possible placements for a segment of a
campaign is the number of possible places (e.g., websites, audio
streams, video streams, etc.) to present an advertisement to the
demographic and/or criterion associated with that segment.
Similarly stated, the number of possible placements for a segment
of a campaign is the amount of placement inventory for the
specified criterion of a segment. In some embodiments, the number
of possible placements for each segment associated with a campaign
can be calculated by a placement estimator module (e.g., placement
estimator module 204, shown and described with respect to FIG.
2).
[0073] A weight is calculated for each segment associated with the
campaign, at 410. The weight for a segment is calculated based on
the campaign score, the number of possible placements for the
campaign, and the number of possible placements for that segment.
In some embodiments, a weight module (e.g., weight module 206 shown
and described with respect to FIG. 2) can calculate the weight of
each segment of the campaign. In some embodiments, the weight of a
segment is based at least partially on a relationship (e.g., ratio)
between the number of possible placements for that segment of the
campaign and the number of possible placements for the remaining
segments of the campaign. In other embodiments, the weight can be
based at least partially on a relationship between the number of
possible placements for that segment of the campaign and the number
of possible placements for the entire campaign. In still other
embodiments, any other relationship between the campaign score, the
number of possible placements for the campaign, and the number of
possible placements for the segment can be used to calculate the
weight.
[0074] The weight of each segment is compared to a first threshold
T.sub.1, at 412. In some embodiments, the first threshold T.sub.1
can be a random threshold, a predetermined threshold, an average of
the weights of the segments of a campaign, and/or the like. If the
weight of a segment is less than the first threshold T.sub.1,
advertisements associated with the campaign are not placed at the
segment, at 418. A weight less than the threshold indicates that
that segment of the campaign does not include enough placements to
make placing advertisements associated with the campaign efficient
with respect to spending the budget of the campaign.
[0075] If the weight of a segment is greater than the first
threshold T.sub.1, a performance score of advertisements placed at
that segment is compared to a second threshold T.sub.2, at 416.
Prior to comparing the performance score with the second threshold
T.sub.2, the performance score of each segment associated with the
campaign is calculated, at 414. In some embodiments, the
performance score of each segment can be calculated at a
performance module (e.g., performance module 208 shown and
described with respect to FIG. 2). In some embodiments, the
performance score of a segment can be based at least in part on a
relationship (e.g., a ratio) between a success metric of
advertisements from a campaign placed at that segment and the
number of impressions associated with the campaign for that
segment.
[0076] If the performance score is greater than the second
threshold T.sub.2, an advertisement associated with the campaign
can be placed at the segment, at 422. In some embodiments, prior to
placing advertisements at the segment, each segment with a weight
greater than the first threshold T.sub.1 and a performance score
greater than the second threshold T.sub.2 is ordered based on
performance score. Advertisements are first placed at the
placements associated with the segments having the highest
performance score. Advertisements are then placed at the placements
associated with the other segments until the campaign budget is
spent. Accordingly, the advertisements are placed at the placements
of the segments having the highest performance scores prior to
being placed at the placements of the segments having lower
performance scores.
[0077] In some embodiments, if the performance score is less than
the second threshold T.sub.2, the campaign score is compared with a
third threshold T.sub.3, at 420. The third threshold T.sub.3 can
act as a gate to allow advertisements to be placed at placements
associated with a segment having a performance score less than the
threshold. The third threshold T.sub.3 can be associated with a
time left to spend the budget. For example, the third threshold
T.sub.3 can be based on a number of seconds left in a time to spend
the budget. If the budget score is greater than the third threshold
T.sub.3, there is a large number advertisement units remaining in
the budget and/or a small number of possible placements for the
advertisements of a campaign without much time left to spend the
budget. Accordingly, advertisements are placed at placements of
segments having a performance score of less than the second
threshold T.sub.2, at 422, in order to spend the budget prior to
the time to spend the budget elapsing.
[0078] In other embodiments, any number of thresholds can be used.
In some embodiments, for example, only one of the thresholds
(T.sub.1, T.sub.2, T.sub.3) is used. In other embodiments, for
example, two of the three thresholds (T.sub.1, T.sub.2, T.sub.3)
are used. In yet other embodiments, additional thresholds can be
used.
[0079] While shown and described above as using both a performance
score for a segment and a weight for a segment to determine whether
or not to place advertisements at placements associated with that
segment, in other embodiments, only one of the performance scores
for the segment or the weight for the segment is used to determine
whether or not to place advertisements at placements associated
with the segment. FIG. 6, for example, is a flow chart illustrating
another method 500 of optimizing the allocation of advertisements,
according to another embodiment. The method 500 uses the weight of
the segments of an advertisement campaign to optimize the placement
of advertisements.
[0080] The method 500 includes calculating a budget score for an
advertisement campaign, at 502. The budget score is based on a
number of advertisement units in a campaign budget of the
advertisement campaign and a number of potential placements
matching a criterion of the advertisement campaign. In some
embodiments, the budget score can be calculated similar to the
budget score calculated using the budget module 202, shown and
described with respect to FIG. 2.
[0081] A weight is calculated for each segment from a set of
segments of the potential placements matching the criterion, at
504. The weight for a segment from the set of segments is based on
the budget score and a relationship between a number of potential
placements for that segment from the set of segments and a number
of potential placements for the remaining segments from the set of
segments. In some embodiments, such a relationship can be a ratio
between the number of potential placements for that segment and the
number of potential placements for the remaining segments. In other
embodiments, the weight for a segment from the set of segments can
be based on the budget score and a relationship between the number
of potential placements for that segment and the number of
potential placements matching the criterion of the advertisement
campaign.
[0082] At least one advertisement from the advertisement campaign
is presented to a placement associated with the segment from the
set of segments if the weight for that segment is greater than a
threshold, at 506. Accordingly, the weight of each segment from the
set of segments is used to optimize the placement of
advertisements. Because a budget score is used in the calculation
of the weight of each segment, the campaign budget will be factored
into whether or not an advertisement is placed at a placement of
the segment.
[0083] While various embodiments have been described above, it
should be understood that they have been presented by way of
example only, and not limitation. Where methods described above
indicate certain events occurring in certain order, the ordering of
certain events may be modified. Additionally, certain of the events
may be performed concurrently in a parallel process when possible,
as well as performed sequentially as described above.
[0084] While shown and described above as separately comparing the
weight of a segment to a first threshold and the performance score
of the segment to a second threshold, in other embodiments, a
relationship between the weight of the segment and the performance
score of the segment can be defined and/or calculated and compared
to a single threshold. For example, the weight and performance can
be summed, averaged, and/or combined using any suitable method.
After the weight and performance score for the segment are combined
into a single metric, the metric can be compared to a threshold to
determine whether to place an advertisement at a placement
associated with that segment.
[0085] Some embodiments described herein relate to a computer
storage product with a non-transitory computer-readable medium
(also can be referred to as a non-transitory processor-readable
medium) having instructions or computer code thereon for performing
various computer-implemented operations. The computer-readable
medium (or processor-readable medium) is non-transitory in the
sense that it does not include transitory propagating signals per
se (e.g., a propagating electromagnetic wave carrying information
on a transmission medium such as space or a cable). The media and
computer code (also can be referred to as code) may be those
designed and constructed for the specific purpose or purposes.
Examples of computer-readable media include, but are not limited
to: magnetic storage media such as hard disks, floppy disks, and
magnetic tape; optical storage media such as Compact Disc/Digital
Video Discs (CD/DVDs), Compact Disc-Read Only Memories (CD-ROMs),
and holographic devices; magneto-optical storage media such as
optical disks; carrier wave signal processing modules; and hardware
devices that are specially configured to store and execute program
code, such as Application-Specific Integrated Circuits (ASICs),
Programmable Logic Devices (PLDs), Read-Only Memory (ROM) and
Random-Access Memory (RAM) devices.
[0086] Examples of computer code include, but are not limited to,
micro-code or micro-instructions, machine instructions, such as
produced by a compiler, code used to produce a web service, and
files containing higher-level instructions that are executed by a
computer using an interpreter. For example, embodiments may be
implemented using Java, C++, or other programming languages (e.g.,
object-oriented programming languages) and development tools.
Additional examples of computer code include, but are not limited
to, control signals, encrypted code, and compressed code.
[0087] While various embodiments have been described above, it
should be understood that they have been presented by way of
example only, not limitation, and various changes in form and
details may be made. Any portion of the apparatus and/or methods
described herein may be combined in any combination, except
mutually exclusive combinations. The embodiments described herein
can include various combinations and/or sub-combinations of the
functions, components and/or features of the different embodiments
described.
* * * * *