CLICK FRAUD MONITORING BASED ON ADVERTISING TRAFFIC

Abstract

Various techniques of monitoring for click fraud in an online advertising system are disclosed herein. In one embodiment, a web server includes a traffic module configured to detect a request from a client device, an analysis module configured to perform a determination if the detected request includes a client ID. The web server also includes a calculation module configured to determine a count of requests with client ID and a count of requests without client ID based on the determination by the analysis module. The web server further includes a fraud module configured to monitor for click fraud based at least in part on the determined count of requests with client ID and/or count of requests without client ID from the calculation module and the determination by the analysis module.

Description

BACKGROUND

Pay per click (“PPC”) is an Internet advertising model used to direct traffic to advertisers' websites. Based on such a model, webpage publishers display clickable advertisements in exchange for a fee from advertisers when a user clicks on the displayed advertisements. Often, an advertising network (e.g., Microsoft AdCenter®) can act as an intermediary between advertisers and publishers. The advertising network charges an advertiser for a click on its advertisement and pays the publisher who displays the advertisement.

One problem with PPC advertising is click fraud. Click fraud generally refers to the practice of a person (or a computer program) imitates a legitimate user's clicking on an advertisement for generating a charge per click without having actual interest in a target of the advertisement. Click fraud can be committed under different situations. For example, a publisher may commit click fraud by clicking on advertisements displayed on his/her own webpages to increase fees collected from an advertiser or advertising network. In another situation, a person or business may commit click fraud by repeatedly clicking on a competitor's advertisements with the intent of exhausting the competitor's advertising budget. Even though various detection and prevention techniques have been developed, click fraud remains a serious problem.

SUMMARY

The present technology is directed to analyzing aspects of advertising traffic in an online advertising system and monitoring click fraud based on results of the analysis. For example, one technique can include determining validity of advertising traffic by analyzing client identification (“client ID”) of the received search requests. For instance, an online advertising system can track a number of requests corresponding to a particular client ID. If the number of requests with the same client ID exceeds a predetermined threshold over a period, the online advertising system can mark the received requests as invalid. In another instance, the online advertising system can compare the client IDs of received requests with those previously observed to correspond to a particular requesting source (e.g., identified by an IP address). If a number of the client IDs for the requesting source has been changing more frequently that a threshold, the online advertising system can also mark the received requests as invalid. Optionally, the online advertising system can also mark the particular requesting source as malicious.

In other examples, the present technology can also analyze anonymous search requests (i.e., requests without client Ds) based on corresponding publisher and/or advertiser identification. For instance, the online advertising system can track a number of requests that have client IDs and those without for a particular publisher (or advertiser). The online advertising system can calculate a ratio and/or percentage of anonymous requests to those with client IDs for the publisher (or advertiser). If the calculated ratio exceeds a corresponding threshold, the online advertising system can raise an alarm, mark the received anonymous requests as potentially invalid, and/or perform other suitable operations.

This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a schematic diagram illustrating a computing framework for an online advertising system in accordance with embodiments of the present technology.

FIG. 1B is a schematic diagram illustrating a computing framework for an online advertising system in accordance with additional embodiments of the present technology.

FIG. 2 is a block diagram showing computing components suitable for the advertising network server of FIG. 1A in accordance with embodiments of the present technology.

FIG. 3 is a block diagram showing software modules suitable for the process component in FIG. 2 in accordance with embodiments of the present technology.

FIG. 4A is a flow diagram illustrating a process for monitoring click fraud in accordance with embodiments of the present technology.

FIG. 4B is a flow diagram illustrating another process for monitoring click fraud in accordance with embodiments of the present technology.

FIG. 5A is a flow diagram illustrating a process for checking traffic based on client identification in accordance with embodiments of the present technology.

FIG. 5B is a flow diagram illustrating a process for checking traffic based on client identification and requesting source address in accordance with embodiments of the present technology.

FIG. 6A is a flow diagram illustrating a process for checking anonymous traffic for a publisher in accordance with embodiments of the present technology.

FIG. 6B is a flow diagram illustrating a process for checking anonymous traffic for an advertiser in accordance with embodiments of the present technology.

DETAILED DESCRIPTION

Various embodiments of online advertising systems, components, modules, routines, and processes for monitoring click fraud are described below. In the following description, example software codes, values, and other specific details are included to provide a thorough understanding of various embodiments of the present technology. A person skilled in the relevant art will also understand that the technology may have additional embodiments. The technology may also be practiced without several of the details of the embodiments described below with reference to FIGS. 1A-6B.

As used herein, the term “client ID” generally refers to information useful for identifying, recognizing, and/or otherwise classifying a user or a web browser. For example, a client ID can include a browser cookie (e.g., an HTTP cookie) used for an originating website to send state information (e.g., a packet of data) to a users browser. The user's browser then returns the state information to the originating site during subsequent visits. In other examples, a client ID can also include a query string embedded in a universal resource locator (“URL”), a hidden web form field, a local shared object (“LSO”), a document object model (“DOM”) property of window.name, HTTP authentication information, and/or other suitable identification information. In further examples, a client ID may also include an IP address, alone or in addition to the foregoing identification information. Also used herein, the term “anonymous” generally refers to a state of lacking a corresponding client ID.

As discussed in the Background section, click fraud is a serious problem facing the PPC advertising model. As a result, advertisers, publishers, and advertising networks have worked to develop various monitoring and/or detection techniques. For example, one technique monitors for abnormally high number of clicks coming from one or a small group of computers based on IP addresses. However, as various detection techniques evolve, click fraud activities have become more sophisticated. Automated programs (e.g., clickbots) can now “spread out” clicks to a large number of computers. As a result, detection techniques based on a single aspect of advertising traffic may not be sufficient. As discussed in more detail below, the present technology includes techniques for analyzing multiple aspects of advertising traffic and monitoring and/or detecting click fraud based on the result of such analysis. Accordingly, detection efficiencies of click fraud may be improved over conventional techniques.

FIG. 1A is a schematic diagram illustrating a computing framework 100 for an online advertising system in accordance with embodiments of the present technology. As shown in FIG. 1A, the computing framework 100 can include an advertising network server 102, a publisher server 104, an advertiser server 106, and a client device 108 (e.g., a personal computer) interconnected by a network 110. In one embodiment, the network 110 can be the Internet. In other embodiments, the network 110 can also include a personal area network, a local area network, a storage area network, a backbone network, a Metropolitan area network, a wide area network, a virtual private network, and/or other suitable types of network. Even though only the foregoing components are illustrated in FIG. 1A, in other embodiments, the computing framework 100 can also include additional servers, client devices, networking devices, and/or other suitable components.

In certain embodiments, the advertising network server 102, the publisher server 104, and the advertiser server 106 may be independent from one another. For example, an advertising network may provide the advertising network server 102. A website publisher may provide the publisher server 104, and a merchant may provide the advertiser's server 106. In other embodiments, at least some of the foregoing servers 102, 104, and 106 may be provided by a single entity. For example, an advertising network (e.g., Microsoft AdCenter®) can provide both the advertising network server 102 and the publisher server 104 when publishing own webpages (e.g., www.bing.com). An advertising network can provide all of the foregoing servers 102, 104, and 106. In further embodiments, the computing framework 100 may have other suitable arrangements.

The advertising network server 102, the publisher server 104, and the advertiser server 106 can individually include a web server, an application server, a catalog server, a communication server, and/or other suitable types of server. As shown in FIG. 1A, the advertising network server 102 can be operatively coupled to a database 103 storing advertisements, advertising campaigns, and/or other suitable advertising information. The database 103 can reside in read-only memory, random-access memory, magnetic disk storage media, optical storage media, flash memory devices, and/or other suitable storage media excluding propagated signals. The database 103 can also store advertising traffic information, as described in more detail below with reference to FIG. 2.

In operation, the advertising network server 102 can supply a user 101 with advertisements stored in the database 103 based on search requests from the user 101. For example, in one embodiment, the user 101 may send a webpage request (e.g., an HTTP request) to the publisher server 104 using the client device 108. In response, the publisher server 104 sends a first webpage 112a to the client device 108 via the network 110. As shown in FIG. 1A, the first webpage 112a can include a search field 114 and certain publisher content 119. The client device 108 then displays the received first webpage 112a to the user 101.

The user 101 can enter a search term (e.g., “car”) in the search field 114 and sends a search request with the term to the publisher server 104. The publisher sever 104 redirects the received search request to the advertising network server 102, which retrieves relevant advertisements from the database 103 based on the search term. The advertising network server 102 then provides the retrieved advertisements to the publisher server 104.

The publisher server 104 then sends a second webpage 112b to the client device 108 in response to the search request. As shown in FIG. 1A, the second webpage 112b can include multiple clickable advertisements 116 (identified individually as AD1116a and AD2116b) with corresponding hyperlinks. The clickable advertisements can include descriptions and/or links to advertisers' websites that are related to the search term (e.g., a car manufacturer, a car dealership, a car buying service, etc.). Two advertisements AD1116a and AD2116b are shown in FIG. 1A for illustration purposes only. In other embodiments, the second webpage 112b can include one or any other suitable number of advertisements.

If the user 101 is interested in at least one of the advertisements 126, the user 101 can then cock on the displayed advertisement 116 for further information. For example, the user 101 may click on the displayed Ad1116a to send a click request to the advertising network server 102. In response, the advertising network server 102 redirects the click request to the advertiser server 106. The advertiser server 106 then provides a third website 112c with corresponding advertiser content 118 to the client device 108.

In other embodiments, the advertising network server 102 can provide the user 101 with advertisements based on information other than the search requests. For example, as shown in FIG. 1B, after the publisher server 104 receives a webpage request from the user 101, the publisher server 104 then sends a search request with the requested content (or a summary thereof) to the advertising network server 102. In response, the advertising network server 102 retrieves relevant advertisements from the database 103 based on the requested content. The advertising network server 102 then provides the retrieved advertisements to the publisher server 104. In another example, the publisher server 104 can provide the requested webpage to the client device 108 along with an embedded advertisement request. The client device 108 then displays the requested webpage and executes the embedded advertisement request to retrieve relevant advertisements in the database 103 from the advertising network server 102.

The publisher server 104 then provides a content webpage 115a to the client device 108. The content webpage 115a includes the requested contents 105 and the advertisements 116. If the user 101 is interested in any of the advertisements 116, the user 101 can click on the displayed advertisements 116. For example, the user 101 may click on the displayed Ad1116a to send a click request to the advertising network server 102. In response, the advertising network server 102 redirects the click request to the advertiser server 106. The advertiser server 106 then provides an advertiser webpage 115b with advertiser content 128 to the client device 108.

In any of the operational arrangements discussed above, the search request and corresponding response are generally referred to herein as an “impression path.” The click request and corresponding response is generally referred to herein as a “click path.” The advertising network server 102 can be configured to monitor and/or detect click fraud based on search and/or click requests along the impression and/or click paths, respectively, as described in more detail below with reference to FIGS. 2-6C.

FIG. 2 is a block diagram showing computing system software components suitable for the advertising network server 102 in FIG. 1A or 1B in accordance with embodiments of the present technology. In FIG. 2 and in other Figures hereinafter, individual software components, modules, and routines may be a computer program, procedure, or process written as source code in C, C++, Java, and/or other suitable programming languages. The computer program, procedure, or process may be compiled into object or machine code and presented for execution by a processor of a personal computer, a network server, a laptop computer, a smart phone, and/or other suitable computing devices. Various implementations of the source and/or object code and associated data may be stored in a computer memory that includes read-only memory, random-access memory, magnetic disk storage media, optical storage media, flash memory devices, and/or other suitable storage media excluding propagated signals.

As shown in FIG. 2, the input component 132 may accept communication traffic data 150, such as search requests and click requests from the user 101 (FIG. 1A or 1B), and communicates the accepted information to other components for further processing. The database component 134 organizes records, including advertising records 142 (e.g., advertisements and advertising campaigns) and traffic records (e.g., client IDs corresponding to search requests), and facilitates storing and retrieving of these records to and from the database 103. Any type of database organization may be utilized, including a flat file system, hierarchical database, relational database, or distributed database, such as provided by a database vendor such as the Microsoft Corporation, Redmond, Wash. The process component 136 analyzes the traffic data 150, and the output component 138 generates output data 152 based on the analyzed traffic data 150. Embodiments of the process component 136 are described in more detail below with reference to FIG. 3.

Even though only particular components of the advertising network server 102 are shown in FIG. 2, in other embodiments, the advertising network server 102 can also include a bi-directional interface, a keyword processor, an index processor, a listing processor, a network load balancer, a redirect processor, and/or other suitable components. In further embodiments, the advertising network server 102 can include multiple interconnected servers configured to perform keyword analysis, indexing, listing, redirecting, and/or other suitable tasks.

FIG. 3 is a block diagram showing software modules 130 suitable for the process component 136 in FIG. 2 in accordance with embodiments of the present technology. As shown in FIG. 3, the process component 136 can include a traffic module 160, an analysis module 162, a fraud module 164, and a calculation module 166 interconnected with one other. Each module may be a computer program, procedure, or routine written as source code in a conventional programming language, or one or more modules may be hardware modules.

The traffic module 160 is configured to identify and analyze communication traffic data 150. For example, the traffic module 160 may monitor communication traffic in HTTP protocol and identify search and/or click requests from the client device 108 or the publisher server 104 (FIG. 1A). The identified requests may then be converted into traffic records 144 and/or other suitable data stored in the database 103. The traffic module 160 may have routines including, for example, comparison, character parsing, or other suitable routines.

The analysis module 162 may be configured to analyze the identified search and/or click requests from the traffic module 160 and to determine various attributes of the identified requests. For example, in one embodiment, the analysis module 162 is configured to determine whether the identified requests include a corresponding client ID. In another embodiment, the analysis module 162 is configured to determine whether the identified requests include a publisher ID, an advertiser ID, a corresponding IP address, and/or other suitable attributes. In further embodiments, the analysis module 162 may be configured to compare a client ID of a request to a client ID of previously observed requests that correspond to a particular IP address. The analysis module 162 can than supply the analysis results to the calculation module 166 for further processing.

The calculation module 166 can include counters, timers, and/or other suitable accumulation routines configured to perform various types of calculations to facilitate operation of other modules. For example, in one embodiment, the calculation module 166 may include a counter configured to track a number of client IDs that correspond to a particular IP address. In another example, the calculation module 166 may include a counter configured to track a number of requests with a particular client ID. In yet another embodiment, the calculation module 166 may include first and second counters configured to track a number of requests with client IDs and anonymous requests for a publisher (or advertiser), respectively. The calculation module 166 may further include a divider configured to determine a ratio and/or percentage of the requests with client IDs and anonymous requests.

The fraud module 164 may be configured to monitor and/or detect click fraud based on inputs from the analysis module 162, the calculation module 166, or other input 154 (e.g., offline manual input). For example, in certain embodiments, the fraud module 164 can include comparison routines configured to compare at least one the following parameters to a corresponding threshold:

• • a number of client IDs corresponding to an IP address
  • a number of requests corresponding to a particular client ID
  • a ratio and/or percentage of the requests with client IDs and anonymous requests for a publisher
  • a ratio and/or percentage of the requests with client Ds and anonymous requests for an advertiserIf any of the comparisons indicate that the corresponding threshold has been exceeded, the fraud module 164 can indicate to the output component 138 that the request is invalid, the IP address is malicious, and/or can perform other suitable operations.

FIG. 4A is a flowchart showing a process 200 for monitoring click fraud in accordance with embodiments of the present technology. Even though the method 200 is described below with reference to the computing framework 100 of FIGS. 1A, 1B and the software components/modules of FIGS. 2 and 3, the method 200 may also be applied in other systems with additional or different hardware and/or software components.

As shown in FIG. 4A, the process 200 includes a block 202 of monitoring communication traffic received at the advertising network server 102 (FIG. 1A). In one embodiment, the advertising network server 102 can monitor and analyze a browser request (e.g., a search or click request) for at least the following attributes:

• • Client ID
  • Publisher ID
  • Advertiser ID
  • IP address
  • Host domainIn other embodiments, the advertising network server 102 can also monitor and/or analyze domain classes, referrers, and/or other suitable attributes of web traffic. In further embodiments, the monitoring and/or analyzing may be performed by other servers, processors, and/or suitable components separate from the advertising network server 102.

The process 200 can include a decision block 204 to determine whether the received request includes a client ID. If the received request includes a client ID (e.g., a cookie), the process 200 proceeds to checking traffic based on client ID at block 206 and recording traffic count with client ID at block 207. An example request with a cookie is shown below for illustration purposes:

GET/spec.html HTTP/1.1

Host: www.example.org

Cookie: cookie1=value1; cookie2=value2

Accept: */*

As shown above, the example request is send to a Host at www.example.org and includes two cookies (i.e., “cookie1” and “cookie2”) with corresponding values (i.e., “value1” and “value2”). In other examples, the received request may include one or any other suitable number of cookies and/or other suitable types of client ID.

If the received request does not include a client ID (e.g., in the example request above, the “Cookie” and “Accept” lines are omitted), the process 200 proceeds to recording anonymous traffic count based on publisher ID, advertiser ID, and/or other suitable attributes of the received request at block 208. In other embodiments, the determination may be based on other suitable criteria.

At block 206, the process 200 includes checking the received request based on a client ID. In one embodiment, the process 200 includes checking a number of requests received in a period of time (e.g., one day) for a particular client ID. If the number of requests exceeds a predetermined threshold, the process 200 may indicate that the received request is invalid, as described in more detail below with reference to FIG. 5A. In another embodiment, the process 200 includes checking a number of client IDs that correspond to a particular IP address in a period (e.g., one day). If the number of client IDs exceeds a predetermined threshold, the process 200 may indicate that the received request is invalid and optionally, may indicate the IP address as malicious, as described in more detail below with reference to FIG. 5B. In other embodiments, the process 200 may check the received request based on the client ID based on other suitable criteria.

At block 207, the process 200 includes recording traffic count with client ID. In one embodiment, the process 200 includes incrementing a publisher ID counter for a particular publisher (e.g., CNN.com) when the received request has a publisher ID identifying the publisher. In another embodiment, the process 200 includes incrementing an advertiser ID counter for a particular advertiser (e.g., Gap.com) when the received request has an advertiser ID identifying the advertiser. In other embodiment, the process 200 may also incrementing an overall ID counter irrespective of the publisher ID and the advertiser ID.

At block 208, the process 200 includes recording anonymous traffic count. In one embodiment, the process 200 includes incrementing an anonymous publisher counter for a particular publisher (e.g., CNN.com) when the received request has a publisher ID identifying the publisher. In another embodiment, the process 200 includes incrementing an anonymous advertiser counter for a particular advertiser (e.g., Gap.com) when the received request has an advertiser ID identifying the advertiser. In other embodiment, the process 200 may also incrementing an overall anonymous counter irrespective of the publisher ID and the advertiser ID.

The process 200 then proceeds to block 210 for checking anonymous traffic. In one embodiment, process 200 includes calculating a ratio and/or percentage of anonymous requests (e.g., based on the anonymous publisher counter) to those with client IDs (e.g., based on the publisher ID counter) for a publisher. In another embodiment, process 200 includes calculating a ratio and/or percentage of anonymous requests (e.g., based on the anonymous advertiser counter) to those with client IDs (e.g., based on the advertiser ID counter) for an advertiser. In other embodiments, process 200 includes calculating an overall ratio and/or percentage of anonymous requests (e.g., based on the overall anonymous counter) to those with client IDs (e.g., based on the overall ID counter). If the calculated ratios and/or percentages exceed corresponding thresholds, the process 200 may raise an alarm, mark the received request as potentially invalid, and/or perform other suitable functions. Several embodiments of checking anonymous traffic are described in more detail below with reference to FIGS. 6A-6C.

The process 200 then includes a decision block 212 to determine whether the process should continue. In one embodiment, the process 200 continues if additional requests are received. In other embodiments, the process 200 may continue based on other suitable conditions. As a result, the process reverts to monitoring communication traffic at block 202. Otherwise, the process ends.

In FIG. 4A, the operations of recording traffic count with client ID at block 207 and checking traffic based on client ID at block 206 are shown as parallel to each other. In other embodiments, the foregoing operations may have other performance sequences. For example, as shown in FIG. 4B, recording traffic count with client ID at block 207 and checking traffic based on client ID at block 206 may be performed in series. In further embodiments, the process 200 may include other suitable operations and/or sequences.

FIG. 5A is a flow diagram illustrating a process 206 for checking traffic based on client ID in accordance with embodiments of the present technology. As described above with reference to FIGS. 4A and 4B, the process 206 may be a subroutine of the process 200 of FIG. 4A or 4B. As shown in FIG. 5A, the process 206 includes increasing (e.g., by one or other suitable increments) a traffic count for the identified client ID at block 220. The process 206 then includes comparing the accumulated traffic count with a predetermined threshold at block 222. In one embodiment, the predetermined threshold may be set based on previously observed accumulated counts over a period. In another embodiment, the predetermined threshold may be set manually. In further embodiments, the predetermined threshold may be set based on a combination of previously observed accumulated counts and other suitable parameters.

The process 206 then includes a decision block 224 to determine if the accumulated traffic count exceeds the predetermined threshold. If the predetermined threshold is not exceeded, the process 206 returns to process 200 in FIG. 4A or 4B. If the predetermined threshold is exceeded, the process 206 includes marking the received request at block 226. In one embodiment, the received request may be marked as invalid. In another embodiment, the received request may be marked as suspicious and requiring further inspection (e.g., by an operator). In further embodiments, the received request may be marked in other suitable fashion. Subsequently, the process 206 returns to process 200 in FIG. 4A or 4B.

FIG. 5B is a flow diagram illustrating another process 206 for checking traffic based on client ID and requesting source address in accordance with embodiments of the present technology. As shown in FIG. 5B, the process 206 includes comparing the client ID of the received request to that previously observed to correspond to the source address (e.g., IP address) of the received request at block 230. The process 206 then includes a decision block 232 to determine if the client ID of the received request is new and/or otherwise different from that previously observed. If the client ID is not new, the process 206 returns to process 200 in FIG. 4A or 4B.

If the client ID is new, the process 206 includes increasing (e.g., by one or other suitable increments) an ID count for the particular source address at block 234. The process 206 then includes comparing the accumulated ID count with a predetermined threshold at block 236. In one embodiment, the predetermined threshold may be set based on previously observed accumulated ID counts over a period. In another embodiment, the predetermined threshold may be set manually. In further embodiments, the predetermined threshold may be set based on a combination of previously observed accumulated ID counts and other suitable parameters.

The process 206 then includes a decision block 238 to determine if the accumulated ID count exceeds the predetermined threshold. If the predetermined threshold is not exceeded, the process 206 returns to process 200 in FIG. 4A or 4B. If the predetermined threshold is exceeded, the process 206 includes marking the received request at block 226, as discussed above with reference to FIG. 5A. Optionally, the process 206 can also include marking the source address as potentially malicious at block 239. Subsequently, the process 206 returns to process 200 in FIG. 4A or 4B.

FIG. 6A is a flow diagram illustrating a process 210 for checking anonymous traffic for a publisher in accordance with embodiments of the present technology. As described above with reference to FIGS. 4A and 4B, the process 210 may be a subroutine of the process 200 of FIG. 4A or 4B. As shown in FIG. 6A, the process 210 includes calculating a ratio and/or percentage of anonymous traffic for a publisher identified by the publisher ID in the received request at block 240. In one embodiment, the ratio and/or percentage may be calculated based on the accumulated values of the publisher ID counter and the anonymous publisher counter. In other embodiments, the ratio and/or percentage may be averaged over a period and/or otherwise calculated.

The process 210 then includes comparing the calculated ratio and/or percentage with a predetermined threshold at block 242. In one embodiment, the predetermined threshold may be set based on previously observed values. In another embodiment, the predetermined threshold may be set manually. In further embodiments, the predetermined threshold may be set based on a combination of previously observed values and other suitable parameters.

The process 210 then includes a decision block 244 to determine if the calculated ratio and/or percentage exceed the predetermined threshold. If the predetermined threshold is not exceeded, the process 210 returns to process 200 in FIG. 4A or 4B. If the predetermined threshold is exceeded, the process 210 includes marking the received request at block 226, as discussed above with reference to FIG. 5A. Subsequently, the process 210 returns to process 200 in FIG. 4A or 4B.

FIG. 6B is a flow diagram illustrating a process 210 for checking anonymous traffic for an advertiser in accordance with embodiments of the present technology. The operations of process 210 shown in FIG. 6B are generally similar to those in FIG. 6A except that the operations are performed based on the advertiser instead of the publisher. As a result, corresponding operations are identified with an apostrophe and are omitted here for clarity. Even though the process 210 is described above based on publisher ID and advertiser ID, in other embodiments, the process 210 can be based on a host domain, subdomain, and/or other suitable identifiers.

Specific embodiments of the technology have been described above for purposes of illustration. However, various modifications may be made without deviating from the foregoing disclosure. In addition, many of the elements of one embodiment may be combined with other embodiments in addition to or in lieu of the elements of the other embodiments. Accordingly, the technology is not limited except as by the appended claims.

Claims

1. A web server, comprising: a traffic module configured to detect a request from a client device;an analysis module configured to perform a determination if the detected request includes a client ID;a calculation module configured to calculate at least one traffic count associated with the detected request based on the determination from the analysis module; anda fraud module configured to monitor for click fraud based at least in part on the calculated at least one traffic count from the calculation module and the determination from the analysis module.

2. The web server of claim 1 wherein: the detected request includes a client ID;the at least one traffic count includes a count for the client ID;the calculation module is configured to increase the count for the client ID;the fraud module is configured to compare the increased count for the client ID with a threshold; andif the increased count exceeds the threshold, the fraud module is configured to mark the detected request as invalid.

3. The web server of claim 1 wherein: the detected request includes a client ID and an IP address;the at least one traffic count includes a client ID count for the IP address;the web server further includes a database storing previously observed client IDs corresponding to the IP address;the analysis module is configured to compare the client ID of the detected request with the previously observed client IDs in the database;the calculation module is configured to increase the client ID count for the IP address if the client ID is different than the previously observed client IDs in the database; andthe fraud module is configured to compare the client ID count for the IP address with a threshold.

4. The web server of claim 3 wherein, if the client ID count exceeds the threshold, the fraud module is configured to mark the detected request as invalid.

5. The web server of claim 3 wherein, if the client ID count exceeds the threshold, the fraud module is configured to mark the IP address as malicious.

6. The web server of claim 1 wherein: the detected request includes a publisher ID but does not include a client ID;the at least one traffic count includes an anonymous publisher count;the calculation module is configured to increase the anonymous publisher count based on the publisher ID; andthe fraud module is configured to check for excessive anonymous traffic to a publisher identified by the publisher ID based on the anonymous publisher count.

7. The web server of claim 1 wherein: the detected request includes a publisher ID but does not include a client ID;the at least one traffic count includes an anonymous publisher count;the calculation module is configured to increase the anonymous publisher count based on the publisher ID and calculating a percentage of anonymous traffic to a publisher identified by the publisher ID; andthe fraud module is configured to check for excessive anonymous traffic to the publisher based on the calculated percentage of anonymous traffic to the publisher.

8. The web server of claim 1 wherein: the detected request includes an advertiser ID but does not include a client ID;the at least one traffic count includes an anonymous advertiser count;the calculation module is configured to increase the anonymous advertiser count based on the advertiser ID; andthe fraud module is configured to check for excessive anonymous traffic to an advertiser identified by the advertiser ID based on the anonymous advertiser count.

9. The web server of claim 1 wherein: the detected request includes an advertiser ID but does not include a client ID;the at least one traffic count includes an anonymous advertiser count;the calculation module is configured to increase the anonymous advertiser count based on the advertiser ID and calculating a percentage of anonymous traffic to an advertiser identified by the advertiser ID; andthe fraud module is configured to check for excessive anonymous traffic to the advertiser based on the calculated percentage of anonymous traffic to the advertiser.

10. The web server of claim 1 wherein: the detected request includes a host domain but does not include a client ID;the at least one traffic count includes an anonymous host count;the calculation module is configured to increase the anonymous host count based on the host domain; andthe fraud module is configured to check for excessive anonymous traffic to an host identified by the host domain based on the anonymous host count.

11. The web server of claim 1 wherein; the detected request includes a host domain but does not include a client ID;the at least one traffic count includes an anonymous host count;the calculation module is configured to increase the anonymous host count based on the host domain and calculating a percentage of anonymous traffic to a host identified by the host domain; andthe fraud module is configured to check for excessive anonymous traffic to the host based on the calculated percentage of anonymous traffic to the host.

12. A computer implemented method, comprising: receiving a browser request from a client device;determining if the browser request includes a client ID;if the browser request includes a client ID, checking for excessive traffic corresponding to the client ID and/or checking for excessive client ID count corresponding to an IP address associated with the browser request; andif the browser request does not include a client ID, checking for excessive anonymous traffic to at least one of a publisher identified by a publisher ID in the request and an advertiser identified by an advertiser ID in the request.

13. The computer implemented method of claim 12 wherein checking for excessive traffic corresponding to the client ID includes: increasing a traffic count for the client ID;comparing the traffic count with a traffic count threshold; andif the traffic count for the client ID exceeds the traffic count threshold, marking the request as invalid.

14. The computer implemented method of claim 12 wherein checking for excessive client ID count for the IP address includes: comparing the client ID with one previously observed to correspond to the IP address;if the client ID is different than the previously observed one, increasing the client ID count for the IP address;comparing the client ID count with a threshold; andif the client ID count for the IP address exceeds the threshold, marking the request as invalid and/or marking the IP address as malicious.

15. The computer implemented method of claim 12 wherein checking for excessive anonymous traffic includes: increasing an anonymous publisher count;calculating a percentage of anonymous traffic to the publisher based on the anonymous publisher count;comparing the percentage of anonymous traffic to the publisher with an anonymous host percentage threshold; andif the percentage of anonymous traffic to the publisher exceeds the anonymous host percentage threshold, marking the request as invalid.

16. A computer implemented method, comprising: receiving a browser request from a client device, the browser request including an IP address and at least one of a publisher ID and an advertiser ID;determining if the browser request includes a cookie;if the browser request includes a cookie, increasing a publisher ID count based on the publisher ID;increasing an advertiser ID count based on the advertiser ID;checking for excessive traffic corresponding to the cookie, including: increasing a traffic count for the cookie;comparing the traffic count with a traffic count threshold; if the traffic count for the cookie exceeds the traffic count threshold, marking the request as invalid;checking for excessive cookie count for the IP address, including:comparing the cookie with one previously observed to correspond to the IP address of the request; increasing a traffic count for the cookie;comparing the traffic count with a traffic count threshold;if the traffic count for the cookie exceeds the traffic count threshold, marking the request as invalid and the IP address as malicious;if the browser request does not have a cookie, increasing an anonymous publisher count based on the publisher ID;increasing an anonymous advertiser count based on the advertiser ID;checking for excessive anonymous traffic to a publisher identified by the publisher ID, including: calculating a percentage of anonymous traffic for the publisher based on the anonymous publisher count and the publisher ID count;comparing the percentage of anonymous traffic for the publisher with a publisher anonymous percentage threshold;if the percentage of anonymous traffic for the publisher exceeds the publisher anonymous percentage threshold, marking the request as invalid;checking for excessive anonymous traffic to an advertiser identified by the advertiser ID, including: calculating a percentage of anonymous traffic for the advertiser based on the anonymous advertiser count and the advertiser ID count;comparing the percentage of anonymous traffic for the advertiser with an advertiser anonymous percentage threshold;if the percentage of anonymous traffic for the advertiser exceeds the advertiser anonymous percentage threshold, marking the request as invalid.

17. The computer implemented method of claim 6 wherein determining if the browser request includes a cookie includes analyzing the browser request for at least one of a cookie, a publisher ID, an advertiser ID, a host name, and an IP address.

18. The computer implemented method of claim 16, further comprising: increasing an overall ID count if the browser request has a cookie;increasing an overall anonymous count if the browser request does not have a cookie;calculating a percentage of anonymous traffic based on the overall anonymous count and the overall ID count;comparing the percentage of anonymous traffic with an anonymous percentage threshold;if the percentage of anonymous traffic exceeds the anonymous percentage threshold, raising an alarm.

19. The computer implemented method of claim 16 wherein calculating the percentage of anonymous traffic for the publisher includes dividing the anonymous publisher count by the publisher ID count.

20. The computer implemented method of claim 16 wherein calculating the percentage of anonymous traffic for the advertiser includes dividing the anonymous advertiser count by the advertiser ID count.