Patent Application
20100023375
The present invention relates generally to inventory allocation. More particularly, the present disclosure relates to techniques for allocating overlapping inventory.
The Internet has become a mass media on par with radio and television. Similar to radio and television content, Internet content is largely supported by advertising dollars. Two of the most common types of advertisements on the Internet are banner advertisements and text link advertisements. Banner advertisements are generally images or animations that are displayed within an Internet web page. Text link advertisements are generally short segments of text that are linked to the advertiser's web site via a hypertext link.
To maximize the impact of Internet advertising (and maximize the advertising fees that may be charged), Internet advertising services such as ad networks display advertisements that are most likely to capture the interest of the web user. An interested web user will read the advertisement and may click on the advertisement to visit a web site associated with the advertisement.
To select the best advertisement for a particular web user, an advertising service may use whatever information is known about the web user. However, the amount of information known about the web user will vary heavily depending upon the circumstances. For example, some web users may have registered with the web site and provided information about themselves while other web users may not have registered with the web site. Some registered web users may have completely filled out their registration forms whereas other registered web users may have only provided the minimal amount of information to complete the registration. Thus, the quality of the various different advertising opportunities will vary.
Since the quality of the advertising opportunities will vary, an Internet advertising service should be careful to use the advertising opportunities in the most optimal manner possible. For example, an advertising opportunity for an anonymous web user is not as valuable as an advertising opportunity for a web user who has registered and provided detailed demographic information. Thus, it would be desirable to be able to optimally allocate the various different advertising opportunities to different advertisers and advertising campaigns.
At the same time, each of the advertisers has an interest in receiving the optimal value of the advertising opportunities for their own advertising campaigns. Unfortunately, the interest of the advertisers to receive the optimal value for the money it pays for its advertising opportunities may conflict with the interest of the Internet advertising service.
Methods and apparatus for allocating inventory are disclosed. In the following description, the inventory allocation schemes refer to the allocation of inventory that includes advertising opportunities. Each of these advertising opportunities may be referred to as an “impression.”
In accordance with one embodiment, the system may maintain a plurality of inventory pools of impressions that are projected to be available during a time period, each of the plurality of inventory pools having a corresponding set of attributes. The system may receive a plurality of requests from advertisers, each of the plurality of requests requesting a number of impressions during the time period that satisfy a corresponding demand profile, the demand profile having a corresponding set of attributes. The system may allocate impressions in the plurality of inventory pools to the plurality of requests by balancing an interest of a seller of the impressions to maximize value of remaining inventory in the plurality of inventory pools against an interest of the advertisers in allocating a representative sample of impressions in each of the plurality of inventory pools that satisfy the demand profile for each of the plurality of requests.
In accordance with another embodiment, a set of attributes may be associated with a set of impressions, where the set of impressions are in an inventory pool or the set of impressions is requested. The set of attributes may define any number and types of characteristics of the impressions or individuals associated with those impressions. For instance, the attributes may include age, location, and/or sex. As another example, the attributes may include behavioral attributes, such as prior purchase history or prior web activity.
In another embodiment, the invention pertains to a device comprising a processor, memory, and a display. The processor and memory are configured to perform one or more of the above described method operations. In another embodiment, the invention pertains to a computer readable storage medium having computer program instructions stored thereon that are arranged to perform one or more of the above described method operations.
These and other features and advantages of the present invention will be presented in more detail in the following specification of the invention and the accompanying figures which illustrate by way of example the principles of the invention.
Reference will now be made in detail to specific embodiments of the invention. Examples of these embodiments are illustrated in the accompanying drawings. While the invention will be described in conjunction with these specific embodiments, it will be understood that it is not intended to limit the invention to these embodiments. On the contrary, it is intended to cover alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be practiced without some or all of these specific details. In other instances, well known process operations have not been described in detail in order not to unnecessarily obscure the present invention.
In the following description, each advertising opportunity may be referred to as an impression. The inventory of impressions may be grouped and organized into a plurality of inventory pools. While each of the plurality of inventory pools is a separate pool, two or more inventory pools may overlap with one another. For instance, an inventory pool of males may overlap an inventory pool of individuals living in the U.S., as well as an inventory pool of individuals between the ages of 20-30. Thus, the inventory may be referred to as “overlapping.”
In order to more accurately group the impressions, the inventory may be grouped into disjoint pools. In other words, the disjoint pools do not overlap with one another. Each of these disjoint pools has an associated set of attributes. For example, a disjoint pool may include males who are between the ages of 20-30 who do not live in the U.S.
An Internet advertising service may receive numerous requests from advertisers. Each of these requests may identify a specific number of impressions that are desired that satisfy a particular set of attributes. For instance, an advertiser may request that the Internet advertising service provide a million impressions that are male. In this example, the Internet advertising service has a number of options available to it, since there are a number of disjoint pools that include impressions that are male. For example, the males may be living in the U.S. or outside the U.S., as well as within various age ranges. As a result, there are a number of different ways that the Internet advertising service may satisfy this request from the different disjoint inventory pools. This process is further complicated by the fact that the Internet advertising service may receive requests from a plurality of advertisers, as well as multiple requests from the same advertiser. As a result, allocating impressions can be a complex process.
In the following description, methods for allocating overlapping inventory are described. However, in order to describe these methods, it is useful to first describe the general process by which advertisements are distributed and impressions are collected.
The Internet has become a mass media that includes web sites, many of which are supported by advertising. Specifically, web site publishers provide interesting content that attracts web site viewers and the web site publishers intersperse paid advertisements into the web pages. The fees from the advertisers compensate the web site publisher for the web site hosting and the costs of creating the interesting content that attracts the web users. Some Internet web site advertisements are banner advertisements including an advertiser-supplied graphical image or animation. Other Internet web site advertisements merely consist of simple short strings of text and hypertext links.
The advertisements within an advertisement-supported web site are often provided to a commercial web site publisher by an Internet advertising service.
A web site served by the web site publisher server 120 may contain web pages viewed by web users interested in the Internet retailer's products or services. Thus, when an Internet user at computer 110 requests a web page published by web site publisher server 120 (along request communication 111), web site publisher server 120 may subsequently request an advertisement from Internet advertisement service (e.g., server) 130 (along request 121) and receive an advertisement for the Internet retailer (along response 123). The web site publisher server 120 may then deliver the requested web page along with the selected advertisement along response 113 such that the user at computer 110 is exposed to an Internet advertisement that advertises the offerings of the Internet retailer server 140.
If the Internet user at computer 110 is sufficiently interested in what is presented within the advertisement, the Internet user may click on the advertisement such that the user will be re-directed to the Internet retailer server 140. The user may be re-directed to the Internet retailer server 140 through an advertising service server 130 that will record the user's selection of the advertisement in order to bill the advertiser for the selection of the advertisement. Once the user has been re-directed to the Internet retailer server 140, the user at computer 110 may purchase goods or services directly from the Internet retailer server 140.
The web site publisher server 120 may have a user registration database 127 wherein web site publisher server 120 stores information about users who have registered with the web site in order to obtain premium content. When a registered web user visits web site publisher server 120, then web site publisher server 120 may provide some demographic information about that registered user to advertising service server 130 along with advertisement request 121. An advertisement selector 135 at advertising service server 130 may use that demographic information about the web user in order to select an advertisement specifically tailored for the demographic group of that web user.
Furthermore, web site publisher server 120 may provide the IP address from the web user's computer 110. In some cases, advertisement selector 135 may use an IP address to geographical location lookup table to determine a general geographic location of the web user's computer 110. Thus, advertisement selector 135 may use that general geographic location when considering what advertisement to select for display to that web user.
Over time advertising service server 130 can store the IP address and/or other demographic information associated with each advertising opportunity. In addition, the browsing history of registered and non-registered users may also be maintained in the form of inferred behavioral attributes. From this information, the advertising service server 130 can statistically determine what proportion of advertising opportunities will have detailed demographic information, what proportion of advertising opportunities will be from a certain geographic location, and other useful advertising forecast information (e.g., sex of the user, age of the user, etc.). That advertising forecast information can be used to optimize the usage of the advertising opportunities. More specifically, the advertising forecast information can be used to allocate these advertising opportunities among requests from advertisers in order to satisfy contracts with the advertisers, as will be described in further detail below.
As shown in
Each of the identified sets may be referred to as disjoint inventory pools, where each of the disjoint inventory pools has an associated set of attributes. Each of these disjoint inventory pools may be used to satisfy requests for advertising inventory, where each of the requests has an associated demand profile. More specifically, the demand profile may have an associated set of attributes that define the requirements of the advertiser for the inventory that will be used to satisfy the requests. For instance, various attributes may identify the sex, age, geographic location, and/or behavioral attributes of the corresponding advertising opportunity.
The sets corresponding to the disjoint inventory pools are used to satisfy requests for advertising inventory having an associated set of attributes. The sets may be labeled as sets Sj={f1j, . . . , fMj}, where fxj is set to one if the set j has an associated attribute corresponding to that particular field, and zero otherwise. For example, the sets illustrated in
Referring to
The various subsets of
The following list of elements can be used to fully define the overlapping inventory optimization problem.
Historical sales information may be used to assign a value to each of the pools of impressions. For instance, a pool that has associated behavioral attributes may be assigned a higher value than a pool that does not have associated behavioral attributes. As another example, a pool that has targeting attributes that indicate an age and sex of the impressions will be more valuable than a pool that merely has a single targeting attribute indicating a sex of the impressions. Values that are assigned may be, for example, integers between 0 and 10.
Historically, inventory has been allocated proportionally. However, this strategy has resulted in substantial revenue loss when advertising campaigns specifically requested inventory that was no longer available. As a result, sellers of inventory have attempted to cherry pick the inventory in order to maximize the value of their remaining inventory. However, this approach has been detrimental to the quality of the inventory that has been guaranteed to advertisers.
As set forth above, the interests of the seller to maximize the value of the remaining inventory will generally motivate the seller to try to retain some of the more valuable inventory. At the same time the interests of the advertisers is to maximize the value of the advertising opportunities that they receive for their advertising dollars. As a result, the interests of the seller conflict with the interests of the advertisers. The disclosed embodiments enable inventory to be allocated fairly to accommodate the interests of the seller of the inventory and the advertisers.
Seller's Interest: Maximizing Remaining Inventory
If we were to maximize the value of the remaining inventory, one way to represent this mathematically would be:
Maximize Σj Vjyj (1)
Stated another way, Sk+1 may designate the set of demands for the surplus supply:
However, this model optimizes inventory allocation for the seller only.
Advertisers' Interest
In order to ensure that the advertisers receive a fair value for their advertising dollars, the inventory may be allocated in a way that each advertiser is allocated a “representative sample” of the inventory. This is also in the long-term interest of the seller since it increases the market value of the seller's inventory.
If every inventory pool could supply every demand, “fairness” could be achieved by proportional allocation. That is, by allocating
X
jk=(sj×dk)/X, where X=Σj,k xjk, the total inventory.
However, in the more general case, every pool will not be able to supply every demand. As a result, we can obtain a solution as close to the representative as possible by maximizing an entropy function:
subject to (2) and (3).
Fairness
In order to fairly allocate inventory, an objective function that is a weighted composite of (1) and (4) may be maximized:
xjk ln(xjk), where γ is a weighting value
subject to (2) and (3). We may think of this as adding a constraint which specifies a fixed value for the entropy, i.e.
for which γ is the Lagrange multiplier, or alternatively of specifying a constraint on the value of the unused inventory, i.e.
where the Lagrange multiplier is γ−1.
We may choose γ as the weighting of the two components. (Alternatively, we may choose a value for the constant specified in (6) or (7).) This weighting value γ may then be applied to find the solution to this problem, as follows.
We define:
cjk={0 j=1, . . . , J; k=1, . . . , K
{Vj j=1, . . . , J; k=K, . . . , 1
then assigning Lagrange multipliers λj and μk to the equations (1) and (2),
ln(xjk)=(λj+μk+cjk)/γ (8)
Substituting this expression back into (1) and (2), the solutions of this problem can be put in the form:
X
jk
=A
j
B
k
s
j
d
k exp(cjk/γ) ∀
where Aj and Bk are given by:
A
j=[Σk|
B
k=[Σj|
In this manner, the interest of the seller may be balanced against the interest of the advertisers using a weighting value. The solution may be obtained by an iterative scaling algorithm, or alternatively, by solving the nonlinear program with a general Natural Language Processing (NLP) algorithm, such as MINOS™, available from Stanford Business Software, Inc. located in Palo Alto, Calif. In this manner, the impressions from the pools of inventory may be allocated (e.g., guaranteed) to the various requests received from advertisers.
In accordance with one embodiment, an advertiser may request a fixed amount of a specific type of inventory. It may then be determined whether this request can be fulfilled, and if so, which inventory from which pools should be allocated to meet this new request, as well as all other requests. This may be accomplished by using the variable dk to identify the requested fixed amount of the specific type of inventory Sk. The solution may then be obtained for Xjk, as set forth above.
Modified Solution
To ensure representative allocation relative to some pre-defined allocation Xjk0, we can modify the entropy term (4) in the composite objective (5) to
For instance, we might choose a cross-entropy function:
This only changes the solution in (8) slightly with ln(xijk) being replaced by ln(xjk)−ln(xjk0). Also, the weighting value γ in (5) can be replaced by γlk, where the multiplier lk is a campaign specific parameter reflecting the importance we give in ensuring representative allocation to the campaign k. The relative weight lk may be derived from domain knowledge and business constraints.
Use of the cross entropy function allows us to relate our approach to the method described above with reference to (5). Writing Δxjk=xjk−xjk0 and f(xjk)=Σj,k xjkln(xjk/xjk0), the Taylor expansion of f(x) about xjk0 is:
F(xjk)=f(xjk0)+Δf′(xjk0)+(Δx2jk/2)f″(xjk0)+ . . .
we note that f(xjk0)=0, f′(xjk0)=1 and f″(xjk0)=1/xjk0. Hence
Referring back to the objective function in (5) ΣjVjyj−γΣj,kln(xjk), we may replace the entropy function
with
which may be replaced with its equivalent
as set forth above. Thus, in this case, we may maximize the value of
(xjk−xjk0)2/2xjk0].
The system may receive a plurality of requests from advertisers at 304, each of the plurality of requests requesting a number of impressions during the time period that satisfy a corresponding demand profile, the demand profile having a corresponding set of attributes. The set of attributes for each demand profile may include at least one of a desired age, gender, or geographical location of the requested number of impressions. Moreover, the set of attributes for each demand profile may indicate desired behavioral attributes of the requested number of impressions. Each of the requests may also indicate a desired position for ad placement within a web page.
The system may allocate impressions in the plurality of inventory pools to the plurality of requests at 306. The system may allocate impressions by balancing an interest of a seller of the impressions to maximize value of remaining inventory in the plurality of inventory pools against an interest of the advertisers in allocating a representative sample of impressions in each of the plurality of inventory pools that satisfy the demand profile for each of the plurality of requests. For instance, the system may calculate a number of impressions in one or more of the plurality of inventory pools to allocate to each of the plurality of requests by applying a weighting value that reflects a weighting of an interest of a seller of the impressions to maximize value of remaining inventory in the plurality of inventory pools against an interest of the advertisers in allocating a representative sample of impressions in each of the plurality of inventory pools that satisfy the demand profile for each of the plurality of requests.
Allocation of impressions may be accomplished by calculating a number of impressions in one or more of the plurality of inventory pools to allocate to each one of the plurality of requests using the expected value of the inventory in each of the plurality of inventory pools that is projected to be available during the time period, the number of impressions projected to be available in each of the plurality of inventory pools during the time period, and a total requested number of impressions for each demand profile. The number of impressions that are projected to be available in each of the plurality of inventory pools during the time period may be ascertained using historical data. The expected value of the inventory in each of the plurality of inventory pools that is projected to be available during the time period may be assigned, as set forth above.
The system may then compute a price to be charged each of the advertisers based upon the allocation of the impressions to the plurality of requests. For instance, each impression may have a charge associated therewith based upon the value assigned to the associated pool of impressions from which the impression was allocated. In this manner, a total price for all of the impressions allocated to a particular request may be determined.
Embodiments of the present invention may be employed to support the operation of an advertising service in any of a wide variety of contexts. For example, as illustrated in
And according to various embodiments, information associated with the user (e.g., attributes) may be obtained using a wide variety of techniques. For example, the information may be obtained via a local application and/or web site and may be accomplished using any of a variety of processes such as those described herein. However, it should be understood that such methods of obtaining this information are merely examples and that the information may be obtained in many other ways.
A web site is represented in
The disclosed techniques of the disclosed embodiments may be implemented in any suitable combination of software and/or hardware system, such as a web-based server used in conjunction with a web-browser. The system of this invention may be specially constructed for the required purposes, or may be a general-purpose computer selectively activated or reconfigured by a computer program and/or data structure stored in the computer. The processes presented herein are not inherently related to any particular computer or other apparatus. In particular, various general-purpose machines may be used with programs written in accordance with the teachings herein, or it may be more convenient to construct a more specialized apparatus to perform the required method steps.
Regardless of the system's configuration, the user's computer 1000, the server 1008, and/or other devices in the network may each employ one or more memories or memory modules configured to store data, program instructions for the general-purpose processing operations and/or the inventive techniques described herein. The program instructions may control the operation of an operating system and/or one or more applications, for example. The memory or memories may also be configured to store data structures such as pools of impressions, software for allocating impressions, various requests received from advertisers, and previously determined allocations of impressions, etc.
Because such information and program instructions may be employed to implement the systems/methods described herein, the disclosed embodiments relate to machine readable media that include program instructions, state information, etc. for performing various operations described herein. Examples of machine-readable media include, but are not limited to, magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CD-ROM disks; magneto-optical media such as floptical disks; and hardware devices that are specially configured to store and perform program instructions, such as read-only memory devices (ROM) and random access memory (RAM). Examples of program instructions include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter.
Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. Therefore, the present embodiments are to be considered as illustrative and not restrictive and the invention is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.
