WO2014026332A1

WO2014026332A1 - Attribution of credit for content item distribution

Info

Publication number: WO2014026332A1
Application number: PCT/CN2012/080150
Authority: WO
Inventors: Juelu ZHANG; Brian S. O'CLAIR; Qifeng TAN; Sean Harvey; Heng LEI; Ling REN
Original assignee: Google Inc.
Priority date: 2012-08-15
Filing date: 2012-08-15
Publication date: 2014-02-20

Abstract

Methods, systems, and apparatus for attribution of credit for content item distribution. In one aspect, a method includes obtaining attribution expressions, each attribution expression including a set of placement term conditions for placement terms, each placement term having a placement term type, and each placement term type having a corresponding priority; for each attribution expression, obtaining a corresponding attribution model, each attribution model specifying a proportional credit distribution to two entities; obtaining an impression record describing an impression of a content item, the impression record including a set of impression placement terms describing a placement of the impression of the content item; selecting attribution expressions having their set of placement term conditions satisfied by the set of impression placement terms; determining a highest priority selected attribution expression; and applying the attribution model of the highest priority selected attribution expression for the impression.

Description

ATTRIBUTION OF CREDIT FOR CONTENT ITEM DISTRIBUTION

BACKGROUND

This specification relates to credit attribution for distribution of a content item.

The Internet provides access to a wide variety of resources, for example, webpages, images, audio files, and videos. Such access to these resources has likewise enabled opportunities for providing relevant additional content. For example, resources of particular interest to a user can be supplemented with content items, such as videos, images, and advertisements. By identifying the relevant content of a resource requested by a user, it is possible to provide relevant content items to the user.

In order to facilitate distribute advertisements and other content items to users, advertisers and other content providers often enter into agreements with content distributors, whereby the content distributor is paid to place the advertisements and/or other content items within resources that are provided to users. A content distributor can place the content items in resources of its own, or the content distributor may use partners to distribute the content items with the partner's resources.

In the case of advertisements, for example, content distributors are often paid by the advertisers and other content providers for each impression of a content item. If the content distributor places the content items through a partner, the partner may be entitled to a portion of the credit, or revenue, that the content distributor receives from the content provider. If the content distributor places an advertisement within another content item, the provider of that content item may also be entitled to a portion of the credit, or revenue, that the content distributor receives from the advertiser.

Content providers, partners, and content distributors often enter into agreements that define the relative distribution of credit between the parties for each impression of a content item. However, due to differences in relative cost to a partner for a placement of a content item and relative value to a content distributor (or content provider) for the placement, the agreements between the parties can become very complex and disjointed, or overly simplified, and accurately determining the actual cost or value of each impression of a content item can become complicated. Accordingly, attribution of credit to the respective parties for each impression of a content item may fail to be appropriately apportioned. SUMMARY

This specification describes technologies relating to attribution of credit for content item distribution.

In general, one innovative aspect of the subject matter described in this specification can be embodied in methods that include the actions of obtaining attribution expressions, each attribution expression including a set of placement term conditions for placement terms, each placement term being one of a plurality of placement term types, and each placement term type having a corresponding priority that differs from the corresponding priority of each other placement term type; for each attribution expression, obtaining a corresponding attribution model associated with the attribution expression, each attribution model specifying a proportional credit distribution to two entities, the proportional credit distribution being for an attributed impression; obtaining an impression record describing an impression of a content item, the impression record including a set of impression placement terms, each impression placement term being one of the plurality of placement term types, and the set of impression placement terms describing a placement of the impression of the content item; selecting attribution expressions, each selected attribution expression having its set of placement term conditions satisfied by the set of impression placement terms;

determining, based on the corresponding priorities of the placement terms of each selected attribution expression, a highest priority selected attribution expression from the selected attribution expressions; and applying the attribution model corresponding to the highest priority selected attribution expression for the impression. Other embodiments of this aspect include corresponding systems, apparatus, and computer programs, configured to perform the actions of the methods, encoded on computer storage devices.

These and other embodiments can each optionally include one or more of the following features.

The plurality of placement term types may include at least one or more of: a partner class term type having conditions that indicate a class of partners associated with a particular attribution expression; a partner term type having conditions that indicate a particular partner associated with a particular attribution expression; an inventory unit term type having conditions that indicate a portion of a website for containing a content item^a line item priority term type having conditions that indicate a level of priority of a content item; a line item environment term type having conditions that indicate an environment in which a content item is served; an advertiser term type having conditions that indicate a particular advertiser associated with a particular content item; a campaign term type having conditions that indicate a campaign associated with a content item; a line item term type having conditions that indicate a particular content item; a content identifier term type having conditions that indicate a particular context in which a content item is served; and a geography term type having conditions that indicate a geographic location associated with an impression of a content item.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for the selected attribution expression; and determining that the highest priority selected attribution expression is the selected attribution expression having a highest priority placement term that has a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that one of the two or more selected attribution expressions includes a geography term; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions having the geography term.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more selected attribution expressions include a geography term; determining that one of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the geography terms included in the two or more selected attribution

expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the geography term that has the placement term condition that has the highest priority. Geography terms that indicate a more specific geographic location may have a higher priority than geography terms that indicate a less specific geographic location.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more expressions do not include a geography term; determining that one of the two or more selected attribution expressions includes a second highest priority placement term from the set of placement terms for that selected attribution expression, wherein the second highest priority placement term is the placement tenn with a highest priority relative to the priority of all other placement terms for the two or more selected attribution expressions that does not match any other placement term included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the second highest priority placement term.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that each of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a same priority as all other placement term conditions of the geography terms included in the two or more selected attribution expressions; determining that one of the two or more selected attribution expressions includes a second highest priority placement term from the set of placement terms for that selected attribution expression, wherein the second highest priority placement term is the placement term with a highest priority relative to the priority of all other placement terms for the two or more selected attribution expressions that does not match any other placement term included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the second highest priority placement tenn.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is an inventory unit term; determining that one of the two or more selected attribution expressions includes an inventory unit term that has a placement term condition that has a highest priority relative to the prioiity of all other placement term conditions of the inventory unit terms included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the inventory unit term that has the placement term condition that has the highest priority.

Determining the highest priority selected attribution expression may comprise:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is a partner term; determining that one of the two or more selected attribution expressions includes a partner tenn that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the partner terms included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the partner term that has the placement term condition that has the highest priority.

The methods embodying innovative aspects of the subject matter described in this specification may also include the actions of identifying, for each selected attribution expression, a pseudo-impression counter, the pseudo-impression counter representing a number of times the selected attribution expression has had its set of placement term conditions satisfied by a set of impression placement terms; incrementing, for each selected attribution expression, the pseudo-impression counter; identifying, for the attribution model that corresponds to the highest priority selected attribution expression, two or more tiers, each tier specifying a proportional credit distribution to two entities for the attributed impression, the proportional credit distribution specified by each tier being different from each other proportional credit distribution specified by each other tier identified for the attribution model; and selecting one of the two or more tiers based on the pseudo-impression counter for the highest priority selected attribution expression; wherein applying the attribution model corresponding to the highest priority selected attribution expression comprises applying the proportional credit distribution specified by the selected tier.

Particular embodiments of the subject matter described in this specification can be implemented so as to realize one or more of the following advantages. Credit for an impression of a content item may be appropriately apportioned between the parties responsible for placing the content item. Agreements between the parties offer significant customization options without being overly complex, disjointed, or overly simplified.

Appropriate credit distribution may lead to a more positive partner experience, in that the content items presented with partner resources will earn the partner an appropriately apportioned amount of credit based on the particulars of each impression. Content distributors may be able to monetize advertisements and other content items mode effectively and efficiently by using established attribution expressions and their corresponding attribution models. An end user experience may also be enhanced due to the likelihood of more relevant content items being favored by partners who are proving the content items with resources requested by the user. The details of one or more embodiments of the subject matter described in this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a block diagram of an example environment in which credit is attributed for distribution of a content item.

Fig. 2 is an example data flow depicting credit attribution for distribution of a content item.

Fig. 3 is an example data flow depicting credit attribution for distribution of a content item and pseudo-impression allocation.

Fig. 4 is a flow diagram depicting an example process for attributing credit for distribution of a content item.

Fig. 5 is a flow diagram depicting an example process for applying an attribution model based on pseudo-impressions.

Like reference numbers and designations in the various drawings indicate like elements.

DETAILED DESCRIPTION

Overview

A content distributor may distribute a content item, such as an advertisement, through multiple partners, such as publishers. The revenue earned from the distribution of the content item is shared among the distributor and partners. However, the amount of revenue shared may vary among each partner, and may also vary depending on the particulars of the placement of the content item with a partner. Accordingly, the content distributor has multiple attribution expressions that describe how credit should be awarded for distributing the content item. A priority scheme is used to determine the priority of the attribution expressions. If multiple attribution expressions apply to a given impression of distributed content, credit for the impression is attributed to the expression with the highest priority. If no attribution expression is satisfied, a default attribution model may detennine the proportional credit distribution between the parties. Each attribution expression includes placement term conditions that indicate when the particular attribution expression is satisfied for an impression of a content item. For example, an attribution expression may contain placement term conditions identifying a particular advertisement and a particular advertiser. In order for that attribution expression to be satisfied, an impression of a content item must be an impression of the particular

advertisement from the particular advertiser. Each attribution expression has a corresponding attribution model that specifies how credit for an impression should be distributed between parties involved in the transaction - e.g., a 50/50 distribution between the content distributor and a partner responsible for the impression.

When multiple attribution expressions are satisfied for a particular impression of a content item, the attribution expression that has the highest priority is identified based on a priority associated with each of the placement terms included in the attribution expression. The attribution model corresponding to the highest priority attribution expression is the model which is applied for the particular impression.

The attribution expressions are important to partners because it gives partners a measure of control over how content items are distributed with their resources (or with their content items). For example, a particular partner may operate a website that offers video content to users, such as television shows. When a partner agrees to allow a content distributor to distribute advertisements on the partner's website (or with the partner's content items), the partner may want to specify a different attribution of credit for different advertisements.

For example, the partner may want a higher proportional distribution of credit for a video advertisement placed on its website than for a banner advertisement. As another example, some of the partner's videos may be more popular than others, and the partner may want a higher proportional distribution of credit for an advertisement that is placed with a popular video. Likewise, a partner may want a higher proportional distribution of credit for an advertisement that is targeted to a first demographic, and may accept a lower proportional distribution for an advertisement that is targeted to a second demographic. Accordingly, attribution expressions may be used to determine when different attribution models should apply for a given impression of an advertisement or other content item. Example Operating Environment

Fig. 1 is a block diagram of an example environment 100 in which credit is attributed for distribution of a content item. A computer network 102, such as a local area network (LAN), wide area network (WAN), the Internet, or a combination thereof, connects a content distributor 104 to partners 106, user devices 108, and often advertisers 110 and content providers 112.

Each partner 106 is a provider of resources. A resource is any data that can be provided by a partner 106 over the network 102 and that is associated with a resource address. Resources include HTML pages, word processing documents, and portable document format (PDF) documents, images, video, and feed sources, to name just a few. The resources can include content, such as words, phrases, pictures, and so on, and may include embedded information (such as meta information and hyperlinks) and/or embedded instructions (such as JavaScript scripts).

A user device 108 is an electronic device that is under the control of a user and is capable of requesting and receiving resources over the network 102. Example user devices 108 include personal computers, mobile communication devices, and other devices that can send and receive data over the network 102. A user device 108 typically includes a user application, such as a web browser, to facilitate the sending and receiving of data over the network 102. The web browser can enable a user to display and interact with text, images, videos, music and other information typically located on a web page at a website on the World Wide Web or a local area network.

A content distributor 104 facilitates the distribution of content items with resources. For example, the content distributor 104 may own content items, or obtain content items, such as videos, images, and documents, from content providers 112. The content items may then be distributed (e.g., to user devices) by the content distributor 104 and one or more partners 106. The content distributor 104 may also obtain advertisements for placement in content items, such as video, image, or text advertisements from advertisers 110. The advertisements are a type of content item, and can be inserted into or provided with other content items provided by content providers 112 and distributed (e.g., to user devices) by the content distributor 104 and one or more partners 106. In some implementations, multiple content distributors may be used to facilitate the distribution of content items with resources. For example, an advertisement distributor may be used to provide advertisements, while a separate video content distributor may be used to distribute video content.

An advertiser 110 is a content item provider that provides advertisements.

Advertisements can take many forms, such as video advertisements, image advertisements, text advertisements, and banner advertisements, to name a few. Advertisements can be embedded in resources and/or other content items that are provided to user devices, or they can be stand-alone resources provided separately from other resources. Advertisements often contain embedded information and/or embedded instructions, and interaction with an advertisement may cause a user device to request a resource from the corresponding advertiser. In some implementations, an advertiser may also be a content provider 112 and may provide advertisements directly to one or more partners 106.

A content provider 112 is an entity that provides content items to the content distributor 104, partners 106, and/or user devices 108. For example, a content provider may own a video content item that the content provider wants to distribute on the World Wide Web or other network. The content provider can provide the video content item to a content distributor or partner that will distribute the content item across a network. In some implementations, a content provider can also be a partner 106. In another implementation, a content provider 112 may be treated as a content distributor when it provides content items directly to a partner 106.

The content distributor 104 is in communication with a data store 114 that stores data and logs which correspond to information related to the distribution of content items and advertisements. For example, the data store 114 may contain advertisements and

advertisement campaign data provided by advertisers 110, and other content items provided by content providers 1 12, as well as data related to the distribution of those content items and information detailing relationships between the content distributor 104 and partners 106.

The content distributor 104 can monetize a content item provided by a content provider 112 by distributing advertisements 118 with the content item 116. For example, advertisers 110 may agree to pay the content distributor 104 a fee to have their

advertisements distributed with content items. Content providers 112 may also agree to pay the content distributor 104 a fee to have their content items distributed. A content provider or advertiser may offer a certain sum per impression of a content item or advertisement, and may order n impressions. The content distributor 104 may allow content providers 112 and advertisers 110 to define targeting rules that take into account attributes associated with user devices and/or content items to provide targeted content items and advertisements for the corresponding users. In some implementations, the targeting rules may be used to vary the price a content provider or advertiser pays for an impression, or they may be used to determine bid price in a content item or advertisement auction. Example targeting rules include keyword targeting, in which content providers and/or advertisers provide bids or an adjusted price for keywords that are associated with resources and/or content items being provided, demographic targeting, in which content item providers and/or advertisers provide bids or an adjusted price for impressions of their content items users that belong to particular demographics, or geographic targeting, in which content providers and/or advertisers provide bids or an adjusted price for impressions of their content items and advertisements based on a location associated with a user device.

The content distributor 104 may distribute a content item 116 and advertisements 118 directly to a user device 108 across the network 102, or may provide the content item 1 16 and advertisements 118 to one or more partners 106, which may provide the content item and advertisements to user devices 108. For example, if a content distributor 104 has agreed to provide n impressions for a particular advertiser, the content distributor may use a partner 106 to deliver some or all of the impressions. The advertisements may be provided with content items, such as a video commercial in a video content item, or an image overlay advertisement provided for an image content item. In some implementations, the

advertisements 118 are provided at the time the content item is presented at a user device, or provided periodically while the content item 116 is being presented to the user device 108. By way of example, if the content item is a video, advertisements may be provided (e.g., by a content distributor or partner) every ten minutes, and/or upon request by the user device, while the video is running on the user device.

Content distributors 104 and partners 106 enter into agreements that define how revenue is to be shared for each serving of content items and/or advertisements. For example, if a content distributor earns x dollars for each impression of a particular advertisement, the content distributor may share a portion of the x dollars it receives from the corresponding advertiser with the partner who is responsible for the impression. In some implementations, a default attribution model is defined by the agreement between the content distributor and partner. The default attribution model specifies the proportional credit distribution for an impression provided by the partner - e.g., 50% to the content distributor and 50% to the partner. The default attribution model may be overridden in certain circumstances to allow for flexibility in revenue sharing based on the particulars of each impression. The circumstances in which the default attribution model is overridden are defined by attribution expressions, which will be discussed in further detail in the sections that follow.

When a user device 108 requests a content item 116 from a content distributor 104 or a partner 106, the user device 108 is provided with the requested content item 116 and one or more advertisements 118. In some implementations an impression record is created that describes the impression and identifies, for example, the partner providing the content item and/or advertisements, the particular content item or advertisement served and its corresponding campaign and advertiser, and other data relating to the impression type, such as geographic location associated with the user device receiving the impression, keyword targeted data associated with the impression, and demographic targeting data associated with the impression. The impression record is used to determine if any attribution expressions are applicable, and which attribution model applies to a particular impression of a content item or advertisement.

Credit Attribution

Fig. 2 is an example data flow 200 depicting credit attribution between a partner 106 and content distributor 104 for distribution of a content item. While portions of the following description describe the distribution of an advertisement with another content item, they are also applicable to the distribution of an advertisement or other content item on its own. In particular, the credit attribution schemes described below can be used with appropriate attribution apportionment for the distribution of a particular content item (e.g., an

advertisement) or a combination of content items (e.g., an advertisement and a sponsored video).

The content distributor 104 obtains attribution expressions 202, e.g., retrieved from a data store 114. In some implementations the attribution expressions 202 are defined by an agreement between the partner and content distributor. For example, when the partner agrees to distribute a content item for the content distributor, a default attribution model may dictate how revenue is divided between the two parties. Attribution expressions may specify when a different attribution model will apply for an advertisement impression.

Each attribution expression 202 includes a set of placement term conditions for placement terms. The placement term conditions indicate the conditions that an impression of an advertisement must satisfy in order for the attribution expression to be applicable.

Example placement terms and conditions include the following:

a) A partner term type having conditions that indicate a particular partner associated with a particular attribution expression. For example, the partner term condition [partner = 1] may indicate a partner having an identifier of " 1." In the absence of an explicit partner term, one may be inferred from the existence of other placement term conditions, such as a placement term condition that is only associated with one partner.

b) An inventory unit (IU) term type having conditions that indicate a portion of a website for containing content items. For example, the IU condition [IU = 3] may indicate a location of a placement of an advertisement on a website. In some implementations, the IU term specifies a hierarchical level of a website on which a placement is located. For example, the "home" page of a website may be specified by [IU = "home"], a sub-page of the home page may be specified by [IU ~ "home/sports"], and a sub-page of the sub-page may be specified by [IU = "home/sports/baseball"].

c) A line item priority term type having conditions that indicate a level of priority of a content item. For example, a line item may describe characteristics of an advertisement, such as the actual advertisement to be served, a start and end date for serving, a number of impressions desired, and a priority. The priority indicates a level of importance of a content item impression relative to other content item impressions that the content distributor has available to serve. For example, a line item priority condition [LIP = 1] may indicate a highest priority associated with an advertisement relative to other advertisements.

d) A line item environment term type having conditions that indicate an environment in which a content item is served. Example line item environments include a video, audio, display, and text environments. An example line item environment condition [LIE = video] may indicate the placing of an advertisement in a video content item. e) An advertiser terra type having conditions that indicate a particular advertiser associated with a particular content item. For example, the advertiser term condition

[advertiser = 2] may indicate an advertiser - having an identifier of "2" ~ corresponding to the placement of a particular advertisement.

f) A campaign term type having conditions that indicate a campaign associated with a content item. For example, the campaign term condition [campaign = 1] may indicate a campaign - having an identifier of " 1"— to which a served advertisement belongs.

g) A line item tenn type having conditions that indicate a particular content item. For example, the line item condition [LI = 3] may indicate a particular advertisement having an identifier of "3."

h) A content identifier tenn type having conditions that indicate a particular context in which a content item is served. In some implementations, a context indicates certain details of a particular content item in which an advertisement is served, such as the genre or name of a video or audio content item in which an advertisement is served. For example, the content identifier (CI) condition [CI = 4] may identify the name of a particular video file in which an advertisement was served.

i) A geography term type having conditions that indicate a geographic location associated with an impression of a content item. For example, the geography (geo) term condition [geo = United States] may indicate the United States as a geographic location associated a user device that was served with an advertisement.

j) a partner class term type having conditions that indicate a class of partners associated with a particular attribution expression. Partner class (PC) term conditions may include, for example [PC = "all partners"], [PC = "all video content partners"], or any other suitable partner classification type. For example, the PC term condition [PC = "all advertisers"] may indicate all partners that are also advertisers.

The foregoing placement terms and conditions are examples, and other appropriate placement terms and placement term conditions may be used. In some implementations, one or more particular placement terms may be required for attribution expressions. For example, an inventory unit term type, partner term type, or partner class term type may be required for an attribution expression. Each placement term type has a corresponding priority that is different from the priority of each other placement term type. For example, the content identifier term type may have a higher priority than an advertiser term type. In some implementations, the priority of placement term types may be used to determine which attribution expression should apply to a given impression of an advertisement. For example, if two attribution expressions have their placement term conditions satisfied by a single impression, the attribution expression having the highest priority - based on the priority of its satisfied placement terms - will be credited for the impression.

For each attribution expression 202, the content distributor 104 obtains a

corresponding attribution model 204 that specifies a proportional credit distribution between two or more parties. In some implementations, the attribution model 204 retrieved from a data store 114. Attribution models, like attribution expressions, may be defined by an agreement, such as an agreement between the content distributor 104 and partner 106, or between the content distributor 104 and the content provider. For example, an attribution model may specify that credit for an impression of an advertisement is split 50/50 between the content distributor and a partner responsible for serving the impression. In some implementations, the attribution model may specify proportional credit distribution between a content distributor and a content provider, or between a content distributor, content provider, and a partner.

The content distributor 104 obtains an impression record 206 that describes an impression of a content item. In some implementations, the impression record 206 includes a set of impression placement terms that describes a placement of the impression of a content item, such as an advertisement. For example, the impression record 206 in the example data flow 200 describes an impression of an advertisement with a set of impression placement terms 210 as follows: [partner = 1, IU = 1, LIP = 2, advertiser = 1, contentID - 1, geo = United States]. The foregoing example placement terms 208 indicate an impression of an advertisement, where the placement was through a partner identified as partner "1," the inventory unit (e.g., location of the placement on the partner's webpage) is identified as inventory unit "1," the line item priority indicates a priority of "2," the advertiser associated with the advertisement is identified as advertiser "1," the content identifier for the content item with which the advertisement was placed is identified as contentID "1," and the location of the user device where the impression was served is indicated as "United States."

The impression records can be gathered by a variety of different ways. For example, in some implementations, the impression record 206 is created by the content distributor 104; however, other entities, such as a partner or a third party advertisement distributor may create the impression record. In some implementations, the impression record 206 is received by the content distributor 104 from a partner or third party. In another implementation, the impression record 206 may be retrieved from the data store 114. The set of impression placement terms may include a single placement term, all available placement terms, or a combination of two or more placement terms.

The content distributor 104 selects attribution expressions whose corresponding set of placement terms conditions are satisfied by the set of impression placement terms. An attribution expression is considered satisfied by an impression when each of its placement term conditions are satisfied by an impression record. Satisfaction of a particular placement term condition may require an exact match, or, in some implementations, a placement term condition may be satisfied by a specified range of placement terms. For example, a placement term condition may require that the line item priority of an impression be higher than 3 (e.g., LIP < 3) in order for the attribution expression to be satisfied. In the example data flow 200, the selected attribution expressions 210 are AE1, AE2, and AE3. Each selected attribution expression has its placement term conditions satisfied by the set of impression placement terms 208.

The content distributor 104 determines, based on the corresponding priorities of the placement terms of each selected attribution expression, a highest priority selected attribution expression. The highest priority selected attribution expression is one of the attribution expressions included in the selected attribution expressions 210. For example, the impression placement terms 208 may be listed in the following example ascending order of priority: geography, partner, inventory unit, line item priority, line item environment, advertiser, campaign, line item, and content identifier. In some implementations, the highest priority selected attribution expression is the selected attribution expression that has a placement term with the highest priority relative to the priority of all other placement terms of the other selected attribution expressions. In the example data flow 200, attribution expression AE3 is the highest priority attribution expression from the selected attribution expressions 210, because it contains a placement term condition (contentID) that has a higher priority than the respective highest priority placement term conditions specified by the other selected attribution expressions (AE1 and AE2, which are the "partner" and "advertiser" terms respectively).

In the example described above, the attribution expression AE3 was selected because it had the highest priority placement term "ContentID" satisfied. However, there may be situations in which two attribution expressions that each have the same highest priority placement term satisfied. In such situations, one of the placement terms that is not the highest priority placement term may be used to determine which selected attribution expression has higher priority. For example, consider the following two attribution expressions with the same highest priority placement term:

1) [partner = 1, advertiser = 1, contentID = 1, geo = United States]

2) [partner = 1, campaign = 1, contentID = 1]

Using the example order of placement term priority described above, contentID is the highest priority placement term in each of the above attribution expressions. In some implementations, ties may be broken by identifying the next highest priority placement term in each expression, and comparing their corresponding priorities. For example, attribution expression 2 would "win" the tie because its second-highest priority placement term

(campaign) has a higher priority than the second-highest priority placement term in attribution expression 1 (advertiser).

In some implementations, one or more placement terms may be designated as tiebreaker terms that, in the event of a tie between attribution expressions, will be used to break that tie and determine a "winner." The tie breaking term may be considered independent of priority. For example, if the geography term ("geo") is designated as a tiebreaker term, attribution expression 1 "wins" the above tie because it is the only attribution expression that includes a geography term - even if the priority of the geography term is lower than that of the campaign term of attribution expression 2.

The placement term conditions may also have priorities based on the values that satisfy the placement terms. For example, in situations in which the placement term conditions of a tiebreaker term have a corresponding priority, the highest priority placement term condition of a tiebreaker term may be used to break a tie between attribution expressions. To illustrate, consider the following two attribution expressions with the same highest priority placement term, and the geography term as a tiebreaker term:

1) [partner = 1, advertiser = 1, contentID = 1, geo = United States]

2) [partner = 1, campaign = 1, contentID = 1, geo = California]

Each of the geography term conditions has a corresponding priority. The priority may depend, for example, on the precision of the geography term condition, wherein the more precise term conditions have a higher priority than less precise terms. Accordingly, in the foregoing example, attribution expression 2 would "win" a tiebreak because its geography term condition, "California," is more precise than "United States."

In some implementations other tiebreaker terms may also be used. For example, the inventoiy unit term specifies a portion of a website for containing content items, and the website may be represented by a hierarchical tree. The home page of a website may form the root node of the hierarchical tree, with resources and sub-resources of the home page being child nodes and grandchild nodes, respectively. Each level of the hierarchical tree may correspond to an inventory unit term. For example, IU = 1 may indicate the first level of the hierarchical tree - e.g., the root node, or home page - while IU = 2 may indicate a child node of the root node, and IU = 3 may indicate a grandchild node of the root node, and so on. Each inventory unit term condition may have a corresponding priority such that a more precise inventory unit term has a higher priority than a less precise inventory unit term.

Consider the following two attribution expressions:

1) [partner = 1, IU = 1, advertiser = 1, geo = United States]

2) [partner = 1, IU = 3, advertiser = 1, geo = United States]

In some situations, both attribution expressions may be satisfied by a single impression because the inventory unit term may be inclusive when it is representing values of a hierarchical tree structure. For example, an inventoiy unit term that identifies the first level (e.g., the root node) of the hierarchical tree includes all other levels of the tree, and an inventory unit term that identifies a second level (or lower) of the hierarchical tree includes all lower levels of the tree. In other words, IU = 1 includes IU = 2, which includes IU = 3, and so on. Thus, an impression of an advertisement at the third level (or below) of a hierarchical tree representing a website is also an impression encompassed by the first level of the hierarchical tree. In the example attribution expressions above, attribution expression 2 has a higher priority because its inventory unit term is more precise - e.g., IU = 3 is more precise than IU = 1.

In some situations, attribution expressions will not include a tiebreaker term. In order to avoid an actual tie between attribution expressions, the system may be designed to not allow attribution expressions to be created, evaluated, or stored in a data store if attribution expressions could have the same priority without a tiebreaker term. For example, If a first attribution expression is [partner = 1 and line item = 1], a second attribution expression

[partner = 1 and line item = 1 or 2] may not be allowed, because both attribution expressions could be satisfied by a single impression record, and neither include a tiebreaker term.

Once the highest priority selected attribution expression is determined, the content distributor 104 applies, to the impression corresponding to the impression record 206, the attribution model 212 that corresponds to the highest priority selected attribution expression. In the example data flow 200, attribution expression AE3 is the highest priority selected attribution expression, and has a corresponding attribution model 212. The attribution model 212 is the model that is used to specify the proportional credit distribution between the partner 106 and the content distributor 104 for the advertisement impression that corresponds to the impression record 206. For example, the attribution model 212 may specify that the partner is entitled to 70% of the revenue for serving the advertisement associated with the impression record, while the content distributor is entitled to 30% of the revenue.

While the foregoing description describes the application of an attribution model between a partner and content distributor, it is also applicable to an attribution model between a content provider and content distributor. In some implementations, a content distributor may use two attribution models for a single content item impression, one specifying proportional credit distribution between the content distributor and partner, and another specifying proportional credit distribution between the content distributor and content provider. A separate rule may be used to determine which model applies first.

Pseudo-Impressions and Tiers

In some implementations, multiple tiers may be specified for the attribution model of an attribution expression. Multiple tiers allow for different proportional credit distribution for the same attribution model. For example, in some situations the content distributor and partner may want a 50/50 proportional credit distribution, while in other situations a 60/40 proportional credit distribution may be desired. Each different proportional credit distribution for an attribution model may be associated with a tier, and selection of a tier may depend on one or more values associated with an impression record, partner, and/or attribution expression.

For example, pseudo-impressions may be tracked for each attribution expression and/or partner or other placement term type, and they can be used to determine which tier of the corresponding attribution model should be applied for an impression. A pseudo- impression occurs whenever an attribution expression is selected - e.g., whenever an attribution expression's set of placement term conditions is satisfied for a given impression - even if the selected attribution is not the highest priority attribution expression. In other words, each attribution expression, and in some implementations the partner corresponding to the attribution expression, is credited with a pseudo-impression when its set of placement term conditions is satisfied by an impression of content. For an attribution model with multiple tiers, selection of the tier may depend on the number of pseudo-impressions for the corresponding attribution expression and/or partner.

In implementations where pseudo-impressions are tracked for partners, each impression record may result in a single partner being credited with a pseudo-impression. As described above, if a partner is not specified by an impression record, the partner associated with the impression record may be inferred based on the existence of other placement terms included in the impression record. The partner specified by or inferred from the impression record may be attributed with a pseudo-impression. For example, if an impression record specifies [partner = 1], partner "Γ' may be credited with a pseudo-impression. When applying an attribution model corresponding to a selected attribution expression with the highest priority, e.g., one that specifies [PC = "all partners"], the pseudo-impression counter for partner "1" may be used to select a tier associated with the attribution model.

Fig. 3 is an example data flow 300 depicting credit attribution for distribution of a content item and pseudo-impression allocation. In some implementations, the content distributor 104, in addition to performing the operations described above with reference to Fig. 2, may also identify, for each selected attribution expression, a pseudo-impression counter. Each pseudo-impression counter represents a number of times the corresponding selected attribution expression has had its set of placement term conditions satisfied by a set of impression placement terms. In some implementations, the content distributor identifies a pseudo-impression counter for a partner specified by an impression record.

The content distributor 104 increments the pseudo -impression counter for each selected attribution expression. For example, the selected attribution expressions 302 each have a corresponding pseudo-impression counter that is incremented as a result of their placement term conditions being satisfied by the impression placement terms 208. In some implementations, the content distributor increments the pseudo-impression counter for a partner specified by an impression record.

In some implementations, the content distributor 104 identifies tiers 304 associated with the attribution model corresponding to the highest priority selected attribution expression. Each tier specifies a proportional credit distribution to two entities associated with the impression of a content item. For example, each tier may have a corres onding proportional distribution of credit that is different from the other tiers.

The content distributor 104 selects one of the tiers based on the pseudo-impression counter for the highest priority selected attribution expression. In the example data flow 300, "Tier 1" would be selected if the pseudo-impression counter was less than 100, "Tier 2" would be selected if the pseudo-impression counter was between 100 and 999, and "Tier 3" would be selected if the pseudo-impression counter was greater than 1000. The example pseudo -impression counter indicates 500 pseudo-impressions of the attribution expression "AE3," and accordingly, Tier 2 is selected. In some implementations, a tier is selected based on the pseudo-impression counter for the partner specified by the impression record.

When the content distributor 104 applies an attribution model corresponding to the highest priority selected attribution expression, it applies the proportional credit distribution specified by the selected Tier. In the example data flow 300, "Tier 2" is selected, and the content distributor will apply attribution model 306 "AE3(2)" to specify the credit distribution between the content distributor 104 and partner 106. Flow Diagrams Fig. 4 is a flow diagram depicting an example process 400 for attributing credit for distribution of a content item. The process 400 may be used by a data processing apparatus, such as a content distributor.

The process 400 obtains attribution expressions (402). Each attribution expression includes a set of placement term conditions for placement terms, and each placement term is one of a plurality of placement term types. In some implementations, each placement term type has a corresponding priority that differs from the corresponding priority of each other placement type. An example attribution expression may be [partner = 2, advertiser = 3]. The placement term conditions are "partner = 2" and "advertiser = 3", and the placement term types ("partner" and "advertiser") each have a corresponding priority.

The process 400 obtains, for each attribution expression, a corresponding attribution model associated with the attribution expression (404). In some implementations, each attribution model specifies a proportional credit distribution to two entities for an attributed impression. An example attribution model may be [partner = 0.6, content distributor = 0.4], indicating that, as between the partner and content distributor, credit for an impression that satisfies the corresponding attribution expression will be distributed such that the partner gets 60% of the credit, and the content distributor gets 40% of the credit.

The process 400 obtains an impression record describing an impression of a content item (406). In some implementations, the impression record includes a set of impression placement terms, and each impression placement term is one of the plurality of placement term types. The set of impression placement terms describes a placement of the impression of the content item. An example impression record may be [partner = 2, IU = 2, advertiser = 3, campaign = 1, line item = 2, contentID = 1], which describes the placement of an impression of a content item, such as an advertisement. In some implementations, the partner term may be inferred from other terms included in the impression record if the partner term is not included in the impression record.

The process 400 selects attribution expressions (408). In some implementations, each selected attribution expression has its set of placement term conditions satisfied by the set of impression placement terms. Given the example attribution expression and impression record described above, the attribution expression ([partner = 2, advertiser = 3]) may be selected because its placement term conditions are satisfied by the impression record. For example, the placement term conditions may be satisfied because each placement term condition of the attribution expression matches an impression placement term of the impression record.

The process 400 determines a highest priority selected attribution expression from the selected attribution expressions (410). In some implementations, the highest priority selected attribution expression is determined based on the coiresponding priorities of the placement terms of each selected attribution expression. For example, if the priority of "advertiser" is higher than the priority of "publisher," the example attribution expression above would have a higher priority than a second example attribution expression that was simply [partner = 2].

The process 400 applies the attribution model that corresponds to the highest priority selected attribution expression for the impression (412). The attribution model may be applied to determine an apportionment of credit between the content distributor and partner, or between the content distributor and content provider. For example, if the corresponding attribution model is [partner = 0.6, content distributor = 0.4], that will be the model applied to the impression of the content item described by the impression record.

Fig. 5 is a flow diagram depicting an example process 500 for applying an attribution model based on pseudo-impressions. The process 500 may be used by a data processing apparatus, such as a content distributor, to track pseudo-impressions for an attribution expression and select an attribution model for the attribution expression based on a tiering of the expression.

The process 500 identifies a pseudo-impression counter for each selected attribution expression (502). Each pseudo-impression counter represents a number of times the selected attribution expression has had its set of placement term conditions satisfied by a set of impression placement terms. In some implementations, the process 400 increments the pseudo-impression counter for each selected attribution expression (504).

The process 500 identifies two or more tiers for the attribution model that corresponds to the highest priority selected attribution expression (506). Each tier specifies a proportional credit distribution to two entities for the attributed impression, and the proportional credit distribution specified by each tier may be different from each other proportional credit distribution specified by each other tier. For example, an attribution model may have three tiers, the first tier may specify a first proportional credit distribution (e.g., 50/50, or [advertiser = 0.5, content provider = 0.5]), the second tier may specify a second proportional credit distribution (e.g., 60/40), and the third tier may specify a third proportional credit distribution (e.g., 75/25).

The process 500 selects one of the tiers based on the pseudo-impression counter for the highest priority selected attribution expression (508). For example, each tier may correspond to a range of pseudo-impression values - e.g., a first tier may correspond to 0 to 99 pseudo-impressions, a second tier may coirespond to 100 to 999 pseudo-impressions, and a third tier may correspond to greater than 999 pseudo-impressions. The selected tier may be the tier that corresponds to the number of pseudo-impressions represented by the pseudo- impression counter - e.g., if the pseudo-impression counter was 500, the second tier would be selected.

The process 500 applies the proportional credit distribution specified by the selected tier (510) when applying the attribution model corresponding to the highest priority selected attribution expression. Given the foregoing example, the proportional credit distribution of the second tier (60/40) would be applied.

Additional Implementation Details

Embodiments of the subject matter and the operations described in this specification can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions, encoded on computer storage medium for execution by, or to control the operation of, data processing apparatus.

Alternatively or in addition, the program instructions can be encoded on an

artificially-generated propagated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal, which is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. A computer storage medium can be, or be included in, a computer-readable storage device, a computer-readable storage substrate, a random or serial access memoiy array or device, or a combination of one or more of them. Moreover, while a computer storage medium is not a propagated signal, a computer storage medium can be a source or destination of computer program instructions encoded in an artificially-generated propagated signal. The computer storage medium can also be, or be included in, one or more separate physical components or media (e.g., multiple CDs, disks, or other storage devices).

The operations described in this specification can be implemented as operations performed by a data processing apparatus on data stored on one or more computer-readable storage devices or received from other sources.

The term "data processing apparatus" encompasses all kinds of apparatus, devices, and machines for processing data, including by way of example a programmable processor, a computer, a system on a chip, or multiple ones, or combinations, of the foregoing The apparatus can include special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). The apparatus can also include, in addition to hardware, code that creates an execution environment for the computer program in question, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or a combination of one or more of them. The apparatus and execution environment can realize various different computing model infrastructures, such as web services, distributed computing and grid computing infrastructures.

A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, object, or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network. The processes and logic flows described in this specification can be performed by one or more programmable processors executing one or more computer programs to perform actions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read-only memoiy or a random access memory or both. The essential elements of a computer are a processor for performing actions in accordance with instructions and one or more memoiy devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto-optical disks, or optical disks. However, a computer need not have such devices. Moreover, a computer can be embedded in another device, e.g., a mobile telephone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a Global Positioning System (GPS) receiver, or a portable storage device (e.g., a universal serial bus (USB) flash drive), to name just a few. Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices;

magnetic disks, e.g., internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

To provide for interaction with a user, embodiments of the subject matter described in this specification can be implemented on a computer having a display device, e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for displaying information to the user and a keyboard and a pointing device, e.g., a mouse or a trackball, by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback, e.g., visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including acoustic, speech, or tactile input. In addition, a computer can interact with a user by sending documents to and receiving documents from a device that is used by the user; for example, by sending web pages to a web browser on a user's user device in response to requests received from the web browser.

Embodiments of the subject matter described in this specification can be implemented in a computing system that includes a back-end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front-end component, e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the subject matter described in this specification, or any combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network ("LAN") and a wide area network ("WAN"), an internetwork (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks).

The computing system can include users and servers. A user and server are generally remote from each other and typically interact through a communication network. The relationship of user and server arises by virtue of computer programs running on the respective computers and having a user-server relationship to each other. In some embodiments, a server transmits data (e.g., an HTML page) to a user device (e.g., for purposes of displaying data to and receiving user input from a user interacting with the user device). Data generated at the user device (e.g., a result of the user interaction) can be received from the user device at the server.

While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any inventions or of what may be claimed, but rather as descriptions of features specific to particular embodiments of particular inventions. Certain features that are described in this specification in the context of separate

embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination.

Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system components in the embodiments described above should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

Thus, particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

WHAT IS CLAIMED IS:

1. A computer implemented method, comprising:

obtaining attribution expressions, each attribution expression including a set of placement term conditions for placement terms, each placement term being one of a plurality of placement term types, and each placement term type having a corresponding priority that differs from the corresponding priority of each other placement terra type;

for each attribution expression, obtaining a corresponding attribution model associated with the attribution expression, each attribution model specifying a proportional credit distribution to two entities, the proportional credit distribution being for an attributed impression;

obtaining an impression record describing an impression of a content item, the impression record including a set of impression placement terms, each impression placement term being one of the plurality of placement terra types, and the set of impression placement terms describing a placement of the impression of the content item;

selecting attribution expressions, each selected attribution expression having its set of placement term conditions satisfied by the set of impression placement terms;

determining, based on the corresponding priorities of the placement terms of each selected attribution expression, a highest priority selected attribution expression from the selected attribution expressions; and

applying the attribution model corresponding to the highest priority selected attribution expression for the impression.

2. The method of claim 1, wherein the plurality of placement term types include at least one or more of:

a partner class term type having conditions that indicate a class of partners associated with a particular attribution expression;

a partner tenn type having conditions that indicate a particular partner associated with a particular attribution expression;

an inventory unit tenn type having conditions that indicate a portion of a website for containing a content item; a line item priority term type having conditions that indicate a level of priority of a content item;

a line item environment term type having conditions that indicate an environment in which a content item is served;

an advertiser term type having conditions that indicate a particular advertiser associated with a particular content item;

a campaign term type having conditions that indicate a campaign associated with a content item;

a line item term type having conditions that indicate a particular content item;

a content identifier term type having conditions that indicate a particular context in which a content item is served; and

a geography term type having conditions that indicate a geographic location associated with an impression of a content item.

3. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for the selected attribution expression; and

determining that the highest priority selected attribution expression is the selected attribution expression having a highest priority placement term that has a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions.

4. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression;

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that one of the two or more selected attribution expressions includes a geography term; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions having the geography term.

5. The method of claim 1 , wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more selected attribution expressions include a geography term; determining that one of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the geography terms included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the geography term that has the placement term condition that has the highest priority.

6. The method of claim 5, wherein geography terms that indicate a more specific geographic location have a higher priority than geography terms that indicate a less specific geographic location.

7. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more expressions do not include a geography term;

determining that one of the two or more selected attribution expressions includes a second highest priority placement term from the set of placement terms for that selected attribution expression, wherein the second highest priority placement term is the placement tenn with a highest priority relative to the priority of all other placement terms for the two or more selected attribution expressions that does not match any other placement term included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the second highest priority placement terni.

8. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

detennining that two or more of the selected attribution expressions each have a highest priority placement tenn that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; detennining that each of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a same priority as all other placement tenn conditions of the geography terms included in the two or more selected attribution expressions;

detennining that one of the two or more selected attribution expressions includes a second highest priority placement term from the set of placement terms for that selected attribution expression, wherein the second highest priority placement term is the placement term with a highest priority relative to the priority of all other placement terms for the two or more selected attribution expressions that does not match any other placement tenn included in the two or more selected attribution expressions; and

detennining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the second highest priority placement term.

9. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is an inventory unit term;

determining that one of the two or more selected attribution expressions includes an inventory unit term that has a placement term condition that has a highest priority relative to the priority of all other placement tenn conditions of the inventory unit terms included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the inventory unit term that has the placement term condition that has the highest priority.

10. The method of claim 1, further comprising:

identifying, for each selected attribution expression, a pseudo-impression counter, the pseudo-impression counter representing a number of times the selected attribution expression has had its set of placement tenn conditions satisfied by a set of impression placement terms; incrementing, for each selected attribution expression, the pseudo-impression counter; identifying, for the attribution model that corresponds to the highest priority selected attribution expression, two or more tiers, each tier specifying a proportional credit distribution to two entities for the attributed impression, the proportional credit distribution specified by each tier being different from each other proportional credit distribution specified by each other tier identified for the attribution model; and selecting one of the two or more tiers based on the pseudo-impression counter for the highest priority selected attribution expression;

wherein applying the attribution model corresponding to the highest priority selected attribution expression comprises applying the proportional credit distribution specified by the selected tier.

1 1. The method of claim 1, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement tenn for each of the two or more selected attribution expressions is a partner term;

determining that one of the two or more selected attribution expressions includes a partner term that has a placement tenn condition that has a highest priority relative to the priority of all other placement term conditions of the rtner terms included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the partner term that has the placement tenn condition that has the highest priority.

12. A system comprising:

a data processing apparatus; and

a data store storing instructions that, when executed by the data processing apparatus, cause the data processing apparatus to perform operations comprising:

obtaining attribution expressions, each attribution expression including a set of placement tenn conditions for placement terms, each placement tenn being one of a plurality of placement term types, and each placement term type having a corresponding priority that differs from the corresponding priority of each other placement term type; for each attribution expression, obtaining a corresponding attribution model associated with the attribution expression, each attribution model specifying a proportional credit distribution to two entities, the proportional credit distribution being for an attributed impression;

obtaining an impression record describing an impression of a content item, the impression record including a set of impression placement terms, each impression placement term being one of the plurality of placement term types, and the set of impression placement terms describing a placement of the impression of the content item;

13. The system of claim 12, wherein the plurality of placement term types include at least one or more of:

a partner term type having conditions that indicate a particular partner associated with a particular attribution expression;

an inventory unit term type having conditions that indicate a portion of a website for containing a content item;

a line item priority term type having conditions that indicate a level of priority of a content item;

an advertiser term type having conditions that indicate a particular advertiser associated with a particular content item; a campaign term type having conditions that indicate a campaign associated with a content item;

14. The system of claim 12, wherein determining the highest priority selected attribution expression comprises:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement tenn conditions for the selected attribution expression; and

determining that the highest priority selected attribution expression is the selected attribution expression having a highest priority placement term that has a highest priority relative to the priority of all other highest priority placement tenns for the selected attribution expressions.

15. The system of claim 12, wherein determining the highest priority selected attribution expression comprises:

16. The system of claim 12, wherein detennining the highest priority selected attribution expression comprises: identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression;

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more selected attribution expressions include a geography term; determining that one of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a highest priority relative to the priority of all other placement terra conditions of the geography terms included in the two or more selected attribution expressions; and

17. The system of claim 16, wherein geography terms that indicate a more specific geographic location have a higher priority than geography terms that indicate a less specific geographic location.

18. The system of claim 12, wherein determining the highest priority selected attribution expression comprises:

determining that one of the two or more selected attribution expressions includes a second highest priority placement term from the set of placement terms for that selected attribution expression, wherein the second highest priority placement term is the placement term with a highest priority relative to the priority of all other placement terms for the two or more selected attribution expressions that does not match any other placement term included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the second highest priority placement term.

19. The system of claim 12, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that each of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a same priority as all other placement term conditions of the geography terms included in the two or more selected attribution expressions;

20. The system of claim 12, wherein determining the highest priority selected attribution expression comprises:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is an inventory unit term;

determining that one of the two or more selected attribution expressions includes an inventory unit term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the inventory unit terms included in the two or more selected attribution expressions; and

detennining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the inventory unit term that has the placement term condition that has the highest priority.

21. The system of claim 12, wherein the operations further comprise:

identifying, for each selected attribution expression, a pseudo-impression counter, the pseudo-impression counter representing a number of times the selected attribution expression has had its set of placement term conditions satisfied by a set of impression placement terms; incrementing, for each selected attribution expression, the pseudo-impression counter; identifying, for the attribution model that corresponds to the highest priority selected attribution expression, two or more tiers, each tier specifying a proportional credit distribution to two entities for the attributed impression, the proportional credit distribution specified by each tier being different from each other proportional credit distribution specified by each other tier identified for the attribution model; and

selecting one of the two or more tiers based on the pseudo-impression counter for the highest priority selected attribution expression;

22. The system of claim 12, wherein detennining the highest priority selected attribution expression comprises: identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression;

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is a partner tenn;

determining that one of the two or more selected attribution expressions includes a partner term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the partner terms included in the two or more selected attribution expressions; and

determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the partner term that has the placement term condition that has the highest priority.

23. A computer-readable medium encoded with instructions that, when executed by a data processing apparatus, cause the data processing apparatus to perform operations comprising:

obtaining attribution expressions, each attribution expression including a set of placement tenn conditions for placement terms, each placement term being one of a plurality of placement term types, and each placement term type having a conesponding priority that differs from the corresponding priority of each other placement term type;

selecting attribution expressions, each selected attribution expression having its set of placement term conditions satisfied by the set of impression placement terms; determining, based on the corresponding priorities of the placement terms of each selected attribution expression, a highest priority selected attribution expression from the selected attribution expressions; and

24. The computer-readable medium of claim 23, wherein the plurality of placement term types include at least one or more of:

25. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises: identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for the selected attribution expression; and

26. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises:

27. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions and that each of the two or more selected attribution expressions include a geography term; determining that one of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the geography terms included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the geography term that has the placement term condition that has the highest priority.

28. The computer-readable medium of claim 27, wherein geography terms that indicate a more specific geographic location have a higher priority than geography terms that indicate a less specific geographic location.

29. The computer-readable medium of claim 23 , wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of ail other highest priority placement terms for the selected attribution expressions and that each of the two or more expressions do not include a geography term;

30. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises:

identifying, for each selected attribution expression, a highest priority placement term from the set of placement term conditions for that selected attribution expression; determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that each of the two or more selected attribution expressions includes a geography term that has a placement term condition that has a same priority as all other placement term conditions of the geography terms included in the two or more selected attribution expressions;

31. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises:

determining that one of the two or more selected attribution expressions includes an inventoiy unit term that has a placement term condition that has a highest priority relative to the priority of all other placement term conditions of the inventory unit terms included in the two or more selected attribution expressions; and determining that the highest priority selected attribution expression is the one of the two or more selected attribution expressions including the inventory unit term that has the placement term condition that has the highest priority.

32. The computer-readable medium of claim 23 , wherein the operations further comprise: identifying, for each selected attribution expression, a pseudo-impression counter, the pseudo-impression counter representing a number of times the selected attribution expression has had its set of placement term conditions satisfied by a set of impression placement terms; incrementing, for each selected attribution expression, the pseudo-impression counter; identifying, for the attribution model that corresponds to the highest priority selected attribution expression, two or more tiers, each tier specifying a proportional credit distribution to two entities for the attributed impression, the proportional credit distribution specified by each tier being different from each other proportional credit distribution specified by each other tier identified for the attribution model; and

33. The computer-readable medium of claim 23, wherein determining the highest priority selected attribution expression comprises:

determining that two or more of the selected attribution expressions each have a highest priority placement term that has a priority equal to a highest priority relative to the priority of all other highest priority placement terms for the selected attribution expressions; determining that the highest priority placement term for each of the two or more selected attribution expressions is a partner term;