WO2007055812A2 - System and method for targeted ad delivery - Google Patents

System and method for targeted ad delivery Download PDF

Info

Publication number
WO2007055812A2
WO2007055812A2 PCT/US2006/037482 US2006037482W WO2007055812A2 WO 2007055812 A2 WO2007055812 A2 WO 2007055812A2 US 2006037482 W US2006037482 W US 2006037482W WO 2007055812 A2 WO2007055812 A2 WO 2007055812A2
Authority
WO
WIPO (PCT)
Prior art keywords
user
interest score
content category
webpage
identifier
Prior art date
Application number
PCT/US2006/037482
Other languages
French (fr)
Other versions
WO2007055812A3 (en
Inventor
Yiqing Wang
Original Assignee
Disney Enterprises, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Disney Enterprises, Inc. filed Critical Disney Enterprises, Inc.
Publication of WO2007055812A2 publication Critical patent/WO2007055812A2/en
Publication of WO2007055812A3 publication Critical patent/WO2007055812A3/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising

Definitions

  • the present disclosure relates to ad delivery methods and systems. In particular, it relates to delivering personalized ads to online users based on user behavior.
  • Internet advertising has become a prime direct marketing channel for companies selling goods and services. Furthermore, Internet advertising offers the unprecedented opportunity to tailor individualized ads to online users because online users generate behavioral and demographic data that provide insight into their goals and preferences. However, Internet advertising is moving well beyond targeting the user solely by demographic data or by the content of the page being viewed, and is currently being focused on consumers' online behavior. User behavior can be established by tracking user activities Qnline,_ . storing user preferences, etc. Internet advertising, however, is in need of more innovative and effective ways to generate user behavior patterns.
  • Behavior targeting goes beyond reach, frequency and monetary value of a user and can addresses the goal of online advertising which delivers a relevant ad to an interested user at the right time. Behavioral targeting looks at the past behavior of the users and targets advertisements accordingly. Behavioral targeting can be seen as a basic business practice that allows marketers to segment their audiences into manageable groups, deliver the right message to the right person at the right time, and better manage the relationship between marketers and their customers. Behavior targeting uses integrated data from user's online activities to create a comprehensive profile that can be targeted using delivery mechanisms.
  • a webpage identifier that identifies a webpage is generated.
  • the webpage identifier can be a universal resource locator.
  • a content category identifier that identifies a content category is generated.
  • the webpage identifier is associated _ with _ the content category identifier if web content depicted in the webpage is relevant to the content category.
  • a web server then receives a request for the webpage, the request including the webpage identifier and a first user identifier, the request being received from a client computer connected to the computer network.
  • the webpage request can include a webpage request timestamp.
  • a user profile is located by determining whether the first user identifier matches a second user identifier corresponding to the user profile.
  • the user profile includes at least one user content category identifier and a corresponding user interest score.
  • a predetermined amount is established independently from the web content depicted in the webpage being viewed by the user.
  • the corresponding user interest score is increased by the predetermined amount if the content category identifier matches the at least one user content category identifier.
  • the corresponding user interest score is indicative of a level of interest of the user in a content category identified by the at least one user content category identifier.
  • the corresponding user interest score can also be reduced according to a decay rate.
  • the corresponding user interest score is reduced daily regardless of whether a webpage request has been received at the web server.
  • the corresponding user interest score is reduced periodically or from time-to-time regardless of whether a webpage request has been received at the web server.
  • the corresponding user interest score can be reduced according to a long-term decay rate which slowly reduces the corresponding user interest score.
  • the corresponding user interest score can be reduced by a long- term decay rate if the corresponding user interest score surpasses a high threshold.
  • the corresponding user interest score can also be reduced according to a short-term decay rate which rapidly reduces the " corresponding user interest score.
  • the " corresponding User interest score is reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold.
  • the corresponding user interest score can be further reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score, wherein if the corresponding user interest score surpasses a high threshold the corresponding user interest score ceases to be reduced by the short-term decay and instead is reduced by a long-term decay rate.
  • a plurality of decay rates can be used to decay the corresponding user interest score, wherein a decay rate from the plurality of decay rates is chosen to be utilized depending upon the user behavior.
  • the corresponding user interest score of a first content category in the user profile can be decayed at a different rate from the corresponding user interest score of a second content category, wherein the difference is due to the user behaving differently with respect to the first content category in comparison to the second content category.
  • the corresponding user interest score of the first content category in the user profile is decayed slowly if the user shows a long-term interest in the content category.
  • the user can be deemed to show a long-term interest if the user makes webpages requests frequent enough such that the corresponding user interest score increases beyond a predefined threshold.
  • the corresponding user interest score of the second content category in the user profile can be decayed quickly if the user shows a short-term interest in the content category.
  • the corresponding user interest score can be assigned to be reduced by a long-term decay rate which if the corresponding user interest score of the content category identifier surpasses a high threshold. It is then determined whether the user has a long-term interest in the content category by checking whether the content category is assigned to-be reduced-by a long-term decay rate. If it is determined that the user has a long-term interest in the content category, the corresponding user interest score is decayed utilizing the long-term decay rate, wherein the long-term decay rate slowly reduces the corresponding user interest score. If it is determined that the user does not have a long-term interest in the content category, the corresponding user interest score is decayed utilizing the short-term decay rate.
  • the user profile can be deleted if a webpage request is not received from a user after a threshold time.
  • the user profile can be deleted if the corresponding user interest score decreases beyond a low threshold.
  • the content category identifier and the corresponding user interest score can be deleted from the user profile if the corresponding user interest score decreases beyond a low threshold.
  • the corresponding user interest score can be further increased by taking into account an elapsed time between the time of a last webpage request and the time of the webpage request received.
  • the corresponding user interest score can be increased in inversely proportion to the length of time of the elapsed time.
  • the predetermined amount can be utilized to increase the corresponding user interest score every time the user makes a request to a website that has an associated content category. Alternatively, the predetermined amount can be utilized to increase all of the corresponding user interest score.
  • the corresponding user interest score can be a numerical value in a range from zero to one hundred. The predetermined amount can be ten points.
  • the webpage identifier, and the user identifier, in a log file are logged in a log file.
  • the webpage identifier can be further associated with a second content category identifier that identifies a second content category, such that the when the webpage is requested, a second user interest score corresponding to the second content category identifier is increased by the predetermined amount.
  • the user profile is implemented utilizing a data, structure comprising a content category field, an interest score-field, and a timestamp field.
  • the user profile can be for example a database record, a map, or an instance of a class.
  • there is a system to deliver advertisement over a computer network comprising a merchant server and a user profile database.
  • the merchant server is can be configured to receive a webpage request from a web browser on the user computer, the webpage request including a universal resource locator and a user identifier.
  • the user profile database can be configured to store user profiles, each user profile comprising a plurality of content category identifiers, wherein each of the plurality of content category identifier has a corresponding user interest score and a timestamp indicating when the webpage request is last received. The corresponding user interest score increases based on the frequency a webpage associated with the content category is requested.
  • the corresponding user interest score can be indicative of a level of interest of the user in a content category identified by the at least one user content category identifier.
  • the user interest can also be reduced according to a decay rate.
  • the user interest score can be reduced daily regardless of whether a webpage request has been received at the web server.
  • the user interest score can be reduced according to a long-term decay rate which slowly reduces the corresponding user interest score.
  • the corresponding user interest score can be reduced by a long-term decay rate if the corresponding user interest score surpasses a high threshold.
  • the corresponding user interest score can also be reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score.
  • the corresponding user interest score can be reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold.
  • a client computer transmits a request for a webpage, the request including a webpage identifier and a first user identifier.
  • the request being received by a web server that locates a user profile by determining whether the first user identifier matches a second user identifier corresponding to the user profile.
  • the web server further locates a content category identifier in the user profile by determining whether web content depicted in the webpage is associated to the content identified by the content category identifier.
  • the web server increases a corresponding user interest score by a predetermined amount that is independent from the web content depicted in the webpage.
  • the requested webpage is then received from the web server at the client computer.
  • the webpage received at the client computer can include Internet content wherein the Internet content is related to a content category in the user profile having an associated interest score higher than a predetermined amount.
  • the Internet content can include advertisement, music, or video.
  • Figure 1 illustrates a component diagram of an ad delivery system for an Internet website.
  • Figure 2 illustrates a flow diagram of a process in which a user computer accesses the Internet website for the first time.
  • Figure 3 illustrates a flow diagram of a process in which a user computer accesses the Internet website subsequent to the first time access.
  • Figure 4A illustrates content categories and web pages associated to the content categories.
  • Figure 4B illustrates the data structure of a user profile.
  • Figure 4C illustrates a user profile and associated short-term and long-term decay rates.
  • Figure 4D illustrates a user profile and associated short-term and long-term decay rates.
  • Figure 5A illustrates a decay curve wherein the decay rate is a short-term decay rate.
  • Figure 5B illustrates a decay curve wherein the decay rate is a long-term decay rate.
  • the method and system described herein establishes the level of interest that a user has in a given content category and delivers ads or content based on such interest level.
  • a user interest is defined as the user's preference in the content categories.
  • the interest is the result of a human being's behavior pattern which drives the user back to collect information in the same content category.
  • the interest can be quantified by a number to be used to compare and select among the users. Unlike other systems, in which all demographic and behavioral information is recorded, the current system simply stores a numerical value that represents the user interest.
  • the overall profile of a user can be established by assigning an interest score for each content category.
  • Each interest score varies depending on the frequency of the visits and the amount of time in between visits to any webpage within an content category.
  • ad serving platforms and content management systems can present ads or adjust content of the ad depending on the interest score of each visitor.
  • the variability of the interest score permits merchants to use this system to efficiently take advantage of fleeting revenue opportunities.
  • a user profile includes multiple interest scores, each interest score representing user interest for a particular content category. Moreover, each interest score can be increased according to a high frequency of visits, or decreased depending on decay. A short-term decay or a long-term decay rate can be applied to the interest score depending on the user behavior after over a period of time.
  • Figure 1 illustrates a component diagram of an ad delivery system for an
  • a merchant server 108 provides information, sells merchant products, and advertises according to user behavior.
  • the merchant server 108 includes a web server application 110 which retrieves a requested web page from the web pages database 112.
  • the web server application also conducts online sales transactions such as providing a product catalog stored in a product catalog database 134 and storing buyers' information in an orders and buyers database 132.
  • the web server application 110 interacts with an ad delivery application 114 which provides ad information to the web server application 110.
  • a content delivery application can also be used to interact with the web server application 110 and deliver targeted content to a user.
  • the web server application 110 receives a request through the Internet 106, the web pages sent back are routed to the user's computer 100.
  • the user's computer 100 utilizes a web browser 102 to interface through the Internet
  • the web browser 102 has access to a cookies database 104 where cookies received from web server application 110 can be stored.
  • the web browser 102 can also interact with the cookies database 104 to transmit the cookie message back to the web server application 110 when a web page associated with the cookie is requested.
  • a logging server 130 provides capabilities of logging and " recording data regarding user webpage requests.
  • the logging server 130 includes a logging application 128 that interacts with the web server application 110 in order to receive user webpage request information and store such information on a log files database 124.
  • the log files database 124 includes hash tables indexed by a user identifier.
  • a scoring application 126 can be located at the logging server 130 and can utilize webpage request information received from logging application 128 in order to calculate interest scores.
  • the scoring application 128 utilizes algorithms that emulate the human memory function.
  • the scoring application 128 can be configured with algorithms that calculate short-term interest scores and long-term interest scores. Short-term interest scores emulate short-term memory, while long-term interest score emulate long-term memory.
  • the scoring application 126 populates interest score data on a user profile database 122 located on a data access server 116.
  • the data access server 116 also includes a cleanup application 118 and a reporting application 120.
  • the cleanup application 118 and the reporting application 120 can be configured as scheduled jobs within a scheduling framework. When running together, the cleanup application 118 should be scheduled prior to the reporting application 120. Both the cleanup application 118 and the reporting application 120 recursively read, process, and update all the user profiles.
  • the data access server 116 and logging server 130 can be combined to be on one single server.
  • the logging application 128, the scoring application 126, the clean up application 118, and the reporting application 120 can run in simultaneously on the single server, as well as the user profile database
  • FIG. 2 illustrates- a flow diagram of a " process 200 in which a user " computer accesses the Internet website for the first time.
  • process block 210 a webpage request is received from the user's computer 100.
  • the web server application 110 receives a request through the Internet 106.
  • the web server application 110 can assume that the web browser 102 is accessing the website hosted by the merchant server 108 for the first time.
  • a user identifier is generated by the web server application 110. If the web server application 110 determines that the web browser 102 is accessing the website for the first time, a user identifier is uniquely assigned to the user's computer 100.
  • the user identifier can be the internet protocol (IP) address of the user's computer 100.
  • the user identifier is generated by the logging application 128. Because the logging application 128 can store the user identifiers in a log file in the log files database 124, the logging application 128 can generate a user identifier based on the previously assigned user identifiers.
  • a cookie is implanted on the user's computer 100.
  • the web server application 110 After generation of the user identifier, the web server application 110 packages the user identifier into a cookie and sends the cookie to the web browser 102. In subsequent times that the user accesses the website hosted by the web server application 110, the web browser can also send the user identifier which permits the ad delivery system to target personalized ads to the user.
  • the user identifier, the request time, and the universal resource locator requested are stored in the log files database 124.
  • the user identifier, the time at which the webpage was received by the web server 110, and the universal resource locator are hashed to a swapping hash table.
  • the user identifier can be used as the key for the hash table.
  • a new hash table can be created for each user and be stored as part of the log files database 124.
  • the request time can be used as the key for the user hash table.
  • the user's activities can be combined together and ordered by request time.
  • the scoring application 126 populates the user profile database 122 with the calculated interest score, content category, and request time.
  • the log files database 124 is periodically processed by the scoring application 126 to update the user profile database 122.
  • the scoring application 126 can be configured to look up the user identifier in the user profile database 122 in order to locate the user profile and adjust the interest scores of the user.
  • the scoring application 126 follows the model, and emulates the function of the human cerebellum, a brain subdivision whose main function is associative memory. Like the cerebellum, which makes strong neural associations for information that is constantly revisited, the scoring application 126 stores the user interest and increases the interest score according to an increase on the website visit frequency. The user interest score for each content category is increased according to the frequency of webpage requests. Thus, more webpage requests by a user in a certain amount of time, the higher the user interest for that particular content category. Moreover, like memory associations that are not constantly revisited in the brain, if the user does not visit a website frequently enough, the score continually decreases until the interest score reaches a low threshold level.
  • the scoring application 126 populates the user profile database 122 every time a new entry is made on the log file database 124. Thus, every time a user requests a -webpage, the scoring application 126 updates the interest scores for that user. For each visit by each user, the scoring application 126 updates the user profile and sends the new profile data to the user profile database 122. In another embodiment, the scoring application 126 populates the user profile database 122 periodically or at any regular or irregular time interval.
  • the scoring application 126 updates the user profile database 122 to reduce all the interest scores in the user profile database 122.
  • the scoring application 126 updates the interest scores depending on either a short-term interest formula or a long-term interest formula. Factors such as the frequency of webpage requests and time delay in between webpage requests can equally be applied by both the short-term interest formula and the long-term interest formula.
  • the short-term interest formula can have a decay rate higher than the decay rate of the long-term interest formula.
  • the long-term interest of a user on a particular content category can be reflected based on the decay rate.
  • the cleanup application 118 disposes of redundant or useless user profiles in the user profile database 122.
  • a user profile is considered no longer useful if the user profile has not been modified for a long period of time. If the user profile has not been modified for a long period of time, the cleanup application 118 can assume that the user has not visited the website for a long time. Through this mechanism, the cleanup application 118 ensures that user profiles with long-term interest decay rates are eliminated if the user profile has been stored for a period of time beyond a threshold level.
  • a user profile is considered no longer useful if each interest score in the user profile is below a predefined threshold level. For example, where the interest score decays rapidly, the user profile may be eliminated unless the user visits the website very frequently. This mechanism can help the cleanup application 118 to ensure " that the user profile database 122 keeps user profiles of users genuinely interested in short-term content.
  • the reporting application 120 queries the user profile database 122 and reports the user profile information.
  • the reporting application 120 queries the user profile database 122 and utilizes the extracted information to forecast inventory, adjust ad and content delivery strategy, or retrieve any information on the user profiles.
  • FIG. 220 the webpage requested by the user is transmitted from the web server application 110 to the web browser 102.
  • the web server application 110 retrieves the necessary data from the webpages database 112 in order to transmit the requested webpage.
  • Figure 3 illustrates a flow diagram of a process 300 in which a user computer accesses the Internet website subsequent to the first time access.
  • a webpage request and a cookie message are received from the user's computer 100.
  • the user's computer already has an implanted cookie on the cookies database 104.
  • the user's computer 100 automatically generates a webpage request with the message indicated by the cookie, namely the user identifier.
  • the web server application 110 receives the webpage request from the web browser 102, a cookie accompanies the webpage request which indicates to the web server application 110 that the user is a returning user.
  • process block 216 Much like the first time a user requests the webpage, subsequent requests of the webpage triggers the above-mentioned of storing of webpage request information as shown in process block 216, populating the user profile database 218 as shown in process block 218, and performing cleanup and reporting operations as shown in process blocks 22 and 224, respectively. Therefore, as stated above, in process block 216 the user identifier, the request time, and the universal resource locator requested are stored in the log files database 124. In process block 218, the scoring application 126 can populate the user profile database 122 with a recalculated interest score. In addition, in process block 222 unused user profiles are eliminated, and in process block 224 user profiles can be outputted on a report format.
  • a user profile is retrieved from the user profile database 122 by the ad delivery application 114.
  • the web server 110 sends the user identifier to the ad delivery application 114.
  • the ad delivery application 114 utilizes the user identifier to query the user profile database 122 and obtain the user profile corresponding to the user requesting the webpage. With the user profile available, the ad delivery application 114 can generate an ad tailored to the interests of the user.
  • an ad is generated based on the user profile retrieved from the user profile database 122.
  • the ad delivery application 114 retrieves user profiles directly from the user profile database 122.
  • the user profile binary information can be decoded by a user segmentation application layer and later sent to the ad delivery application 114 in a format compatible with the ad delivery application 114.
  • the segmentation application layer decodes the user profile and makes the user profile available to the ad delivery application 114. Furthermore, other applications located at the merchant server 130 can also retrieve user profiles efficiently without having to access and decode the user profile information from the user profile database 122.
  • the user profile information is decoded dynamically.
  • the user identifier is used to query the segmentation application layer. If the user profile exists in the user profile database 122, the user profile can be decoded and modified to adjust the interest score that indicates the current interest level for each segment according to the current webpage request.
  • a communication protocol between the ad delivery application 114 and the segmentation application layer is applied to transform the profile into a string of data that can be used directly by the ad delivery application 114 to apply the targeting logic.
  • the user profile database 122 can be updated immediately after receiving webpage request such that the user profile retrieved by the ad delivery application 114 includes the latest webpage request. For example, if the user requested a sports webpage, the webpage request information is reflected on the user profile, such that when ad delivery application 114 access the user profile, the user profile includes the webpage request for the sports webpage.
  • the ad delivery application 114 can utilize interest scores that are updated with realtime data so the webpage requested includes behavioral ad targeting that takes into account the current webpage request.
  • the ad delivery application 122 can be further utilized for inventory.
  • An inventory management system can be utilized in tandem with the ad delivery application 122 in order to forecast inventory needs.
  • the interest scores retrieved from the user profile database 122 and the advertisement presented to the users can be computed to calculate the demand for merchant products.
  • the ad delivery application 122 can be used for targeted content delivery.
  • Interest scores can further be used by the web server application 110 to tailor content displayed to the user. By providing personalized content, the user overall experience is enhanced thus increasing the probabilities that the user can continue requesting webpages from the merchant.
  • process block 320 the webpage requested by the user is transmitted from the web server application 110 to the web browser
  • the web server application 110 can include the generated ad as part of the requested webpage.
  • Webpages with similar content can be classified under similar content categories.
  • human behavior is considered as a consistent process, and therefore online users consistently access the same or similar websites.
  • Grouping similar webpages into content categories facilitates the segmentation of user behavior because a user can be assigned an interest score for each content category.
  • Figure 4A illustrates content categories and web pages associated to the content categories.
  • Universal resource locators and corresponding content categories can be stored in a local database in the merchant server 108 or in the logging server 124.
  • the local database includes a relational table of content categories and universal resource locators associated with at least one content category.
  • Each content category can reflect a common area of interest.
  • a website 402 for the Entertainment and Sports Programming Network can have a content category 404 for Golf, a content category 406 for Formula One
  • One or more webpages can be associated with each content category.
  • a webpage 416 related to the Golf US Open can be associated with the content category 404.
  • an identifier for webpage 416 is matched with the content category that had been previously associated with the webpage 416, namely content category 404.
  • a user requesting a webpage 416 about the GoTf US Open indicates a general level of interest in the content category 404 corresponding to Golf.
  • a webpage 418 related to the Golf Players can be matched with the content category 404.
  • the request can be associated with content category 404.
  • a user that is interested in information on Golf Players can be assumed to be interested, at least partially, in Golf.
  • Formula One Cars, and Car Racing News can be assumed to be generally interested in Formula One Car Racing.
  • a webpage 420 that includes content on Formula One Drivers can be associated with content category 406 for
  • a webpage 422 related to the Formula One Cars can be associated with the content category 406.
  • a webpage 424 related to Car Racing News can also be associated with Formula One Car Racing. Then, if a user requests webpage 424, the request is matched with content category 406 reflecting the user interest in Formula One Car Racing.
  • Car Racing can have one or more webpages associated with the content category 408.
  • a webpage 426 that includes content on IndyCar Drivers, and a webpage 428 that includes content on IndyCar Cars can be associated with content category 406 for IndyCar Car Racing.
  • a user that visits the car racing news webpage 424 can also be deemed to be interested in IndyCar Car Racing.
  • a webpage can be associated with one or more content categories.
  • Webpage 424 for example, can be associated with both Formula One Car Racing and IndyCar Car Racing.
  • the user can be assumed to be interested in Formula One Car Racing and IndyCar Car Racing.
  • a user profile is developed for each user.
  • Figure 4B illustrates the data structure of a user profile.
  • Each user profile 430 can be implemented utilizing known data structures such as tables, database records, instances of an object-oriented class (an object), etc.
  • a user profile 430 can be a record of a user profile table 432 located in database 122.
  • a user profile can be an instance of a class having multiple attributes.
  • Each user profile 430 can be uniquely identified by a user identifier.
  • the user profile table 432 can include a plurality of user identifiers and the data for each corresponding user profile.
  • each profile 430 can be a generic map that includes values corresponding to the content category identifier 440, the interest score 436, and the timestamp 438.
  • the content category identifier 440 can be numerical value that is interpreted as a content category.
  • the content category identifier 440 can be character string corresponding to the name of the content category, for example if the content category is Golf, the category identifier 440 can be a character string "Golf.”
  • the interest score 436 is a point system represented by a numerical value with a range of 0 to 100 points.
  • the interest score 436 is a numerical value with a range of 0.00 to 1.00 points.
  • the interest score 436 can have any other numerical value range or be stored as a character string.
  • the timestamp 436 can also be stored as a numerical value as well as a character string.
  • the user profile 430 includes illustrated in Figure 4B, includes three content categories: Golf, Formula One Car Racing, and IndyCar Car Racing.
  • the interest score 436 for Golf is 25 points, and the last request was made at 03/06/2005 at 8:32 AM.
  • the user profile 430 in Figure 4B indicates that the user has in interest level of 15 points for the content category Formula One Car Racing and that the last visit to a webpage associated with the Formula One Car Racing content category was on 8/21/2004 at 7:57 PM.
  • the user category for IndyCar Car Racing shows that the user does not have an interest in this content category.
  • the user profile can simply not include content category in which the user has no interest.
  • a user profile 430 can simply have one entry for Golf as the content category identifier 440, the interest score 436 as 25 points and the timestamp 438 of the last request.
  • the user profile can have five content categories, each with a corresponding interest score 436 and a timestamp 438.
  • the interest score 436 corresponding to each of the content categories can be calculated according to different contributing factors.
  • the factors that can determine the interest score can include 1) the frequency of visits to any of the webpages associated with the content category; 2) the time elapsed from one visit to the next; and 3) a decay rate.
  • the interest score 436 changes based on the frequency of visits to any of the webpages associated with the content category. The more frequent a user visits any of the webpages associated with the content category, the higher the interest score 436. Thus, every time a user visits any of the webpages associated with the content category, the interest score 436 can be increased by a fixed amount.
  • the content category associated the webpage is identified. For example, if a user visits the Formula One Car Racing webpage 428 illustrated in Figure 4A, the system maps the universal resource locator associated with the Formula One Car Racing webpage 428 to the corresponding category Formula One Car Racing 406. Based on the user identifier, the user profile 430 is retrieved and the interest score 436 for Formula One Car Racing is updated.
  • the interest score 436 can be increased by 1, 5, 10 or any other arbitrary setting configured in the system to be the standard increase for a visit to a webpage in the content category.
  • the amount of increase can be configured to be, for example, 10 points every time the user visits a webpage in the content category. In another embodiment, the amount of increase can be 5 points.
  • the interest score 436 for a content category can increase based on the total number of visits to the webpages associated with the content. For example, referring to Figure 4A, visits the webpage 416 about the Golf US Open as well as visits to webpage 418 related to the Golf Players increases the interest score 436 of the content category Golf in the user profile 430. In the user profile
  • the fixed increase can be configured to
  • each of the interest scores in the content categories can be increased.
  • the car racing news webpage 424 is associated with two content categories, namely content category 408 corresponding to IndyCar Car Racing, and the content category 408 corresponding to Formula One Car Racing.
  • the interest score 436 corresponding to each of the content categories increases.
  • the fixed increase is configured to be 10 points
  • the interest score 436 for Formula One Car Racing increases from 15 points to 25 points.
  • the interest score 436 for IndyCar Car Racing will now be 10 points.
  • the interest score 436 changes based on the time elapsed from one visit to the next. While the number of visits over a certain period of time is directly proportional to the interest score 436, the time between visits can also provides with insightful information about user behavior. For example, a first user might visit the Formula One Car Racing webpage four times in a period of twelve weeks. The four visits can occurred within a period of four minutes during the first week. A second user might also visit Formula One Car Racing webpage four times in a period of twelve weeks. However, the second user might visit once a week for the last four weeks of the twelve-week period. While the visitation frequency of both users is the same, namely four visits in twelve weeks, the behavior is not the same.
  • the elapsed time between visits can also be taken into account when calculating the interest score 436 for each category.
  • the user profile 430 corresponding to each user is a closer- prediction of user's- interest for certain content.
  • the elapsed time is calculated by subtracting the current time from the timestamp 438 the last visit.
  • the user profile 430 include the timestamp 438 he last visit to any of the webpages associated with the content category 440. For example, if the content category is Golf, the timestamp 438 can be for the last visit to the webpage 416 related to the Golf US Open. If the user then visits the same webpage 416, two days later, the calculated time elapsed will be two days.
  • the elapsed time will also be two days because webpage 418 and webpage 416 correspond to the same content category Golf.
  • a user that visits a webpage within short intervals of time is assumed to have higher interest in the content of the webpage, and therefore in the content category.
  • the relationship between the interest score and the amount of time elapsed is inversely proportional. The shorter the time elapsed from one visit to the next, the higher the interest score.
  • the elapsed time can be expressed in terms of a multiplier fraction that has the amount of time in the denominator and one in the numerator. The elapsed time can be expressed in number of minutes, hours, etc.
  • the interest score 436 changes based on a decay rate.
  • the interest score 436 is decayed based on a configured decay rate. Users that visit websites of a certain content might not come back to visit for a long time, if ever. The current system therefore applies a decay rate for every user interest score 436. Thus, if the user continues to visit the website the interest score 436 is maintained, but if the user ceases to visit websites of the given content category, the user interest score 436 decreases to the point that the system no longer recognizes the user as being interested in the
  • the decay rate is applied uniformly to all interest scores 436 of all the user profiles.
  • the system can decay interest scores on a regular basis.
  • all interest scores of all user profiles are decayed by a percentage (e.g. 7.5%) everyday.
  • all interest scores of all user profiles are decayed by a percentage (e.g. 15%) every week.
  • the decay is a constant fixed number. For example, if interest score range is 1 to 100 points, the decay rate can be 1 point per day.
  • each interest score is decayed every time the user visits certain category, wherein the decay takes into consideration the amount of time elapsed such that the higher the time elapsed, the higher the decay rate applied.
  • a long-term and short-term decay rates can be used.
  • the user visits webpages associated with a particular content category a high number of times, the user can be assumed to have a long-term interest in that content category.
  • some content categories can have seasonal information that by definition greatly attracts users for short period of time until the next season arrives. If the user makes many visits to webpages related to a specific content category, the user can be deemed to have a long-term interest in that content category.
  • the decay rate applied can be a long-term decay rate that decays the interest score for the seasonal content category much slower.
  • Figure 4C illustrates a user profile and associated short-term decay rates.
  • each user profile in the system can be divided into two portions: " a firsTpofti ⁇ n containing content categories and corresponding interest " scores that decay with a short-term decay rate, and a second portion containing content categories and corresponding interest scores that decay with a long-term decay rate.
  • the user profile 430 can be configured to differentiate between the content categories in which the user has a long-term interest and a short-term interest.
  • each of the content category in a the user profile 430 can have a corresponding attribute or flag (implemented with any data type) that can indicate whether the decay applied to the interest score that particular content category is to be decayed with the short-term decay rate or the long term decay rate.
  • each user profile 430 can include two tables, one for the content categories with long-term interest, and another for the content categories with short-term interest.
  • the user profile 430 can have all of the content categories having a short term applied. Namely, the decay rate attributes
  • the interest score 436 for each category can increase, decrease or remain approximately the same. In order to remain the same, the user must frequently visit webpages associated with the content category. As the decay is applied to the interest score of the category, the decrease due to the decay can be compensated by the increase due to visits to webpages corresponding to the content category.
  • a low threshold can be established such that when the interest score decreases due to decay applied, interest score that reaches the low threshold indicates to the system that the user is no longer interested in the content category corresponding to the low interest score.
  • the content category can then be removed from the content categories in the user profile. If the content category was the only one left in the listing of content categories, the user profile can be removed from the user profile database.
  • a high threshold can be established such that when the interest score increases due very frequent visitation to webpages associated with the content category. When the interest score reaches the high threshold, the system can recognize the user as a "power user" and change the decay rate applied to the content category to be a long-term decay.
  • Figure 4D illustrates a user profile and associated short-term and long-term decay rates.
  • a threshold level for a category "Golf can be established to be 40 points. Any user that reaches 40 points or beyond qualifies as a "power user" on the category of Golf, and therefore the decay rate applied can be a long- term decay rate. If the user has an interest score of 25 points for the Golf content category, the user can visit a Golf-related webpage twice in one day and accumulate 20 extra points (assuming 10 points are awarded per visit).
  • the decay rate applied 442 changes from short-term to long-term.
  • all further application of decay to the interest level of the section golf can be long-term.
  • the decay rate attributes 444 and 446 remain as "Short Term,” reflecting that the user has a short-term interest in Formula One Car Racing and IndyCar Car Racing.
  • the decay rate attribute 442 is now "Long Term" reflecting that the user has a long-term interest in Golf.
  • the decay can be calculated periodically for all of the interest scores in the system.
  • the system applies the long-term decay rate to the interest scores that are flagged to have a long-term decay, and a short-term decay rate to the interest score that are flagged to have a short-term decay. For example, if the short-term decay rate is 20% and the long-term decay rate is 7.5%, all interest scores in all user profiles that are associated with content categories that have a long-term decay rate applied will be decayed by 7.5%. Likewise, all interest scores in all user profiles that are associated with content categories that have a short-term decay rate applied are decayed by 20%.
  • the high threshold level can be the same for all content categories (e.g. 50 points). In another embodiment, the high threshold level can be different for each content category (e.g. 50 points for Golf, and 40 points for Formula One Car Racing).
  • Figure 5A illustrates a decay curve wherein the decay rate is a short-term decay rate.
  • the curve shows the decay of the interest score for a content category in which the decay rate applied is a 20% decay.
  • the initial interest score is 40 points. Without further interaction by the user, such as user visits to webpages related to the content category, the interest score decreases rapidly. For example, point 402 shows that on the seventh day the interest score has decreased down to 10 points.
  • Figure 5B illustrates a decay curve wherein the decay rate is a long-term decay rate.
  • the curve shows the decay of the interest score for a content category in which the decay rate applied is a 7.5% decay rate.
  • the initial interest score is 40 points.
  • the interest score decreases but not as rapidly as with a decay rate of 20% as show in Figure 5A above.
  • point 404 of Figure 5B shows that not until the nineteenth day does the interest score decrease down to 10 points.
  • a user can show interest in a content category not only by viewing associated webpages, by also by accessing web content related to the content category. For example images, video, audio, text, ads, flash animations, and any other web content that is related to the substance of the particular content category.

Abstract

Generating a user profile for advertisement, as shown in Fig. 4B, over a computer network is disclosed. A webpage identifier and a content category are generated. The webpage identifier is associated with the content category identifier if web content depicted in the webpage is relevant to the content category. A web server then receives a request for the webpage, the request including the webpage identifier and a first user identifier. A user profile is located by determining whether the first user identifier matches a second user identifier corresponding to the user profile. The user profile includes at least one user content category identifier and a corresponding user interest score. A predetermined amount is established independently from the web content depicted in the webpage being viewed by the user. The corresponding user interest score is increased by the predetermined amount if the content category identifier matches the user content category identifier.

Description

SYSTEM AND METHOD FOR TARGETED AD DELIVERY
BACKGROUND OF THE DISCLOSURE
1. Field of the Disclosure
The present disclosure relates to ad delivery methods and systems. In particular, it relates to delivering personalized ads to online users based on user behavior.
2. General Background
Internet advertising has become a prime direct marketing channel for companies selling goods and services. Furthermore, Internet advertising offers the unprecedented opportunity to tailor individualized ads to online users because online users generate behavioral and demographic data that provide insight into their goals and preferences. However, Internet advertising is moving well beyond targeting the user solely by demographic data or by the content of the page being viewed, and is currently being focused on consumers' online behavior. User behavior can be established by tracking user activities Qnline,_ .storing user preferences, etc. Internet advertising, however, is in need of more innovative and effective ways to generate user behavior patterns.
Behavior targeting goes beyond reach, frequency and monetary value of a user and can addresses the goal of online advertising which delivers a relevant ad to an interested user at the right time. Behavioral targeting looks at the past behavior of the users and targets advertisements accordingly. Behavioral targeting can be seen as a basic business practice that allows marketers to segment their audiences into manageable groups, deliver the right message to the right person at the right time, and better manage the relationship between marketers and their customers. Behavior targeting uses integrated data from user's online activities to create a comprehensive profile that can be targeted using delivery mechanisms.
Current systems of behavior targeting, however, typically record every user visit to a website, the visit time, and other information that has proven to be excessive because of the associated storage space required to record such information. In other current systems, a user that visits a website a certain threshold number of visits is automatically tagged as having one behavior so no further action is taken to reflect user behavior changes. As such, current systems have not yet exploited the full potential of behavior targeting on Internet advertising.
SUMMARY
In one aspect, there is a method of generating a user profile for advertisement over a computer network. A webpage identifier that identifies a webpage is generated. The webpage identifier can be a universal resource locator. A content category identifier that identifies a content category is generated. The webpage identifier is associated _ with _ the content category identifier if web content depicted in the webpage is relevant to the content category. A web server then receives a request for the webpage, the request including the webpage identifier and a first user identifier, the request being received from a client computer connected to the computer network. The webpage request can include a webpage request timestamp. A user profile is located by determining whether the first user identifier matches a second user identifier corresponding to the user profile. The user profile includes at least one user content category identifier and a corresponding user interest score. A predetermined amount is established independently from the web content depicted in the webpage being viewed by the user. The corresponding user interest score is increased by the predetermined amount if the content category identifier matches the at least one user content category identifier.
In a further aspect, the corresponding user interest score is indicative of a level of interest of the user in a content category identified by the at least one user content category identifier. The corresponding user interest score can also be reduced according to a decay rate. In addition, the corresponding user interest score is reduced daily regardless of whether a webpage request has been received at the web server. In another aspect, the corresponding user interest score is reduced periodically or from time-to-time regardless of whether a webpage request has been received at the web server.
In another aspect, the corresponding user interest score can be reduced according to a long-term decay rate which slowly reduces the corresponding user interest score. The corresponding user interest score can be reduced by a long- term decay rate if the corresponding user interest score surpasses a high threshold. In another aspect, the corresponding user interest score can also be reduced according to a short-term decay rate which rapidly reduces the " corresponding user interest score. The "corresponding User interest score is reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold. In yet another aspect, the corresponding user interest score can be further reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score, wherein if the corresponding user interest score surpasses a high threshold the corresponding user interest score ceases to be reduced by the short-term decay and instead is reduced by a long-term decay rate. A plurality of decay rates can be used to decay the corresponding user interest score, wherein a decay rate from the plurality of decay rates is chosen to be utilized depending upon the user behavior. In another aspect, the corresponding user interest score of a first content category in the user profile can be decayed at a different rate from the corresponding user interest score of a second content category, wherein the difference is due to the user behaving differently with respect to the first content category in comparison to the second content category. The corresponding user interest score of the first content category in the user profile is decayed slowly if the user shows a long-term interest in the content category. The user can be deemed to show a long-term interest if the user makes webpages requests frequent enough such that the corresponding user interest score increases beyond a predefined threshold. The corresponding user interest score of the second content category in the user profile can be decayed quickly if the user shows a short-term interest in the content category.
In a further aspect, the corresponding user interest score can be assigned to be reduced by a long-term decay rate which if the corresponding user interest score of the content category identifier surpasses a high threshold. It is then determined whether the user has a long-term interest in the content category by checking whether the content category is assigned to-be reduced-by a long-term decay rate. If it is determined that the user has a long-term interest in the content category, the corresponding user interest score is decayed utilizing the long-term decay rate, wherein the long-term decay rate slowly reduces the corresponding user interest score. If it is determined that the user does not have a long-term interest in the content category, the corresponding user interest score is decayed utilizing the short-term decay rate. The user profile can be deleted if a webpage request is not received from a user after a threshold time. The user profile can be deleted if the corresponding user interest score decreases beyond a low threshold. The content category identifier and the corresponding user interest score can be deleted from the user profile if the corresponding user interest score decreases beyond a low threshold. In another aspect, the corresponding user interest score can be further increased by taking into account an elapsed time between the time of a last webpage request and the time of the webpage request received. The corresponding user interest score can be increased in inversely proportion to the length of time of the elapsed time. The predetermined amount can be utilized to increase the corresponding user interest score every time the user makes a request to a website that has an associated content category. Alternatively, the predetermined amount can be utilized to increase all of the corresponding user interest score. The corresponding user interest score can be a numerical value in a range from zero to one hundred. The predetermined amount can be ten points.
In a further aspect, the webpage identifier, and the user identifier, in a log file are logged in a log file. The webpage identifier can be further associated with a second content category identifier that identifies a second content category, such that the when the webpage is requested, a second user interest score corresponding to the second content category identifier is increased by the predetermined amount. In another aspect, the user profile is implemented utilizing a data, structure comprising a content category field, an interest score-field, and a timestamp field. The user profile can be for example a database record, a map, or an instance of a class. In one aspect, there is a system to deliver advertisement over a computer network, comprising a merchant server and a user profile database. The merchant server is can be configured to receive a webpage request from a web browser on the user computer, the webpage request including a universal resource locator and a user identifier. The user profile database can be configured to store user profiles, each user profile comprising a plurality of content category identifiers, wherein each of the plurality of content category identifier has a corresponding user interest score and a timestamp indicating when the webpage request is last received. The corresponding user interest score increases based on the frequency a webpage associated with the content category is requested.
In another aspect, the corresponding user interest score can be indicative of a level of interest of the user in a content category identified by the at least one user content category identifier. The user interest can also be reduced according to a decay rate. Further, the user interest score can be reduced daily regardless of whether a webpage request has been received at the web server. In another aspect, the user interest score can be reduced according to a long-term decay rate which slowly reduces the corresponding user interest score. In a further aspect, the corresponding user interest score can be reduced by a long-term decay rate if the corresponding user interest score surpasses a high threshold. The corresponding user interest score can also be reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score. In addition, the corresponding user interest score can be reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold.
In one aspect there is a method of receiving a targeted advertisement over a computer network. A client computer transmits a request for a webpage, the request including a webpage identifier and a first user identifier. The request being received by a web server that locates a user profile by determining whether the first user identifier matches a second user identifier corresponding to the user profile. The web server further locates a content category identifier in the user profile by determining whether web content depicted in the webpage is associated to the content identified by the content category identifier. The web server increases a corresponding user interest score by a predetermined amount that is independent from the web content depicted in the webpage. The requested webpage is then received from the web server at the client computer. The webpage received at the client computer can include Internet content wherein the Internet content is related to a content category in the user profile having an associated interest score higher than a predetermined amount. The Internet content can include advertisement, music, or video.
BRIEF DESCRIPTION OF THE DRAWINGS
By way of example, reference will now be made to the accompanying drawings.
Figure 1 illustrates a component diagram of an ad delivery system for an Internet website. Figure 2 illustrates a flow diagram of a process in which a user computer accesses the Internet website for the first time.
Figure 3 illustrates a flow diagram of a process in which a user computer accesses the Internet website subsequent to the first time access.
Figure 4A illustrates content categories and web pages associated to the content categories.
Figure 4B illustrates the data structure of a user profile.
Figure 4C illustrates a user profile and associated short-term and long-term decay rates.
Figure 4D illustrates a user profile and associated short-term and long-term decay rates.
Figure 5A illustrates a decay curve wherein the decay rate is a short-term decay rate.
Figure 5B illustrates a decay curve wherein the decay rate is a long-term decay rate. DETAILED DESCRIPTION
The method and system described herein establishes the level of interest that a user has in a given content category and delivers ads or content based on such interest level. A user interest is defined as the user's preference in the content categories. Thus, the interest is the result of a human being's behavior pattern which drives the user back to collect information in the same content category. The interest can be quantified by a number to be used to compare and select among the users. Unlike other systems, in which all demographic and behavioral information is recorded, the current system simply stores a numerical value that represents the user interest.
As such, the overall profile of a user can be established by assigning an interest score for each content category. Each interest score varies depending on the frequency of the visits and the amount of time in between visits to any webpage within an content category. Thus, ad serving platforms and content management systems can present ads or adjust content of the ad depending on the interest score of each visitor. Furthermore, the variability of the interest score permits merchants to use this system to efficiently take advantage of fleeting revenue opportunities.
In the system and method described below, a user profile includes multiple interest scores, each interest score representing user interest for a particular content category. Moreover, each interest score can be increased according to a high frequency of visits, or decreased depending on decay. A short-term decay or a long-term decay rate can be applied to the interest score depending on the user behavior after over a period of time. Figure 1 illustrates a component diagram of an ad delivery system for an
Internet website. A merchant server 108 provides information, sells merchant products, and advertises according to user behavior. The merchant server 108 includes a web server application 110 which retrieves a requested web page from the web pages database 112. The web server application also conducts online sales transactions such as providing a product catalog stored in a product catalog database 134 and storing buyers' information in an orders and buyers database 132. Furthermore, the web server application 110 interacts with an ad delivery application 114 which provides ad information to the web server application 110. In another embodiment, a content delivery application can also be used to interact with the web server application 110 and deliver targeted content to a user.
Once the web server application 110 receives a request through the Internet 106, the web pages sent back are routed to the user's computer 100. The user's computer 100 utilizes a web browser 102 to interface through the Internet
106 with the web server application 110. The web browser 102 has access to a cookies database 104 where cookies received from web server application 110 can be stored. The web browser 102 can also interact with the cookies database 104 to transmit the cookie message back to the web server application 110 when a web page associated with the cookie is requested.
~A logging server 130 provides capabilities of logging and" recording data regarding user webpage requests. The logging server 130 includes a logging application 128 that interacts with the web server application 110 in order to receive user webpage request information and store such information on a log files database 124. In one embodiment, the log files database 124 includes hash tables indexed by a user identifier.
In one embodiment, a scoring application 126 can be located at the logging server 130 and can utilize webpage request information received from logging application 128 in order to calculate interest scores. In one embodiment, the scoring application 128 utilizes algorithms that emulate the human memory function. Thus, the scoring application 128 can be configured with algorithms that calculate short-term interest scores and long-term interest scores. Short-term interest scores emulate short-term memory, while long-term interest score emulate long-term memory. Once the interest scores are calculated, the scoring application 126 populates interest score data on a user profile database 122 located on a data access server 116. The data access server 116 also includes a cleanup application 118 and a reporting application 120. The cleanup application 118 and the reporting application 120 can be configured as scheduled jobs within a scheduling framework. When running together, the cleanup application 118 should be scheduled prior to the reporting application 120. Both the cleanup application 118 and the reporting application 120 recursively read, process, and update all the user profiles.
In one embodiment, the data access server 116 and logging server 130 can be combined to be on one single server. The logging application 128, the scoring application 126, the clean up application 118, and the reporting application 120 can run in simultaneously on the single server, as well as the user profile database
122 and the log file database 124 being stored on the same server.
- Figure 2 illustrates- a flow diagram of a" process 200 in which a user" computer accesses the Internet website for the first time. In process block 210, a webpage request is received from the user's computer 100. The web server application 110 receives a request through the Internet 106. In one embodiment, the web server application 110 can assume that the web browser 102 is accessing the website hosted by the merchant server 108 for the first time.
Next, in process block 212 a user identifier is generated by the web server application 110. If the web server application 110 determines that the web browser 102 is accessing the website for the first time, a user identifier is uniquely assigned to the user's computer 100. In one embodiment, the user identifier can be the internet protocol (IP) address of the user's computer 100. In another embodiment, the user identifier is generated by the logging application 128. Because the logging application 128 can store the user identifiers in a log file in the log files database 124, the logging application 128 can generate a user identifier based on the previously assigned user identifiers. In process block 214, a cookie is implanted on the user's computer 100.
After generation of the user identifier, the web server application 110 packages the user identifier into a cookie and sends the cookie to the web browser 102. In subsequent times that the user accesses the website hosted by the web server application 110, the web browser can also send the user identifier which permits the ad delivery system to target personalized ads to the user.
In process block 216, the user identifier, the request time, and the universal resource locator requested are stored in the log files database 124. In one embodiment, the user identifier, the time at which the webpage was received by the web server 110, and the universal resource locator are hashed to a swapping hash table. The user identifier can be used as the key for the hash table.
In another embodiment, a new hash table can be created for each user and be stored as part of the log files database 124. The request time can be used as the key for the user hash table. Thus, the user's activities can be combined together and ordered by request time. In process block 218, the scoring application 126 populates the user profile database 122 with the calculated interest score, content category, and request time. In one embodiment, the log files database 124 is periodically processed by the scoring application 126 to update the user profile database 122. The scoring application 126 can be configured to look up the user identifier in the user profile database 122 in order to locate the user profile and adjust the interest scores of the user. As stated above, the scoring application 126 follows the model, and emulates the function of the human cerebellum, a brain subdivision whose main function is associative memory. Like the cerebellum, which makes strong neural associations for information that is constantly revisited, the scoring application 126 stores the user interest and increases the interest score according to an increase on the website visit frequency. The user interest score for each content category is increased according to the frequency of webpage requests. Thus, more webpage requests by a user in a certain amount of time, the higher the user interest for that particular content category. Moreover, like memory associations that are not constantly revisited in the brain, if the user does not visit a website frequently enough, the score continually decreases until the interest score reaches a low threshold level. If the interest score reaches the low threshold level, the corresponding user profile can be deleted, or at least the content category within the user profile can be removed. In one embodiment, the scoring application 126 populates the user profile database 122 every time a new entry is made on the log file database 124. Thus, every time a user requests a -webpage, the scoring application 126 updates the interest scores for that user. For each visit by each user, the scoring application 126 updates the user profile and sends the new profile data to the user profile database 122. In another embodiment, the scoring application 126 populates the user profile database 122 periodically or at any regular or irregular time interval.
In another embodiment, because the user's interest score for each content category is always decaying, the scoring application 126 updates the user profile database 122 to reduce all the interest scores in the user profile database 122. The scoring application 126 updates the interest scores depending on either a short-term interest formula or a long-term interest formula. Factors such as the frequency of webpage requests and time delay in between webpage requests can equally be applied by both the short-term interest formula and the long-term interest formula. However, the short-term interest formula can have a decay rate higher than the decay rate of the long-term interest formula. Thus, the long-term interest of a user on a particular content category can be reflected based on the decay rate. In process block 222, the cleanup application 118 disposes of redundant or useless user profiles in the user profile database 122. In one embodiment, a user profile is considered no longer useful if the user profile has not been modified for a long period of time. If the user profile has not been modified for a long period of time, the cleanup application 118 can assume that the user has not visited the website for a long time. Through this mechanism, the cleanup application 118 ensures that user profiles with long-term interest decay rates are eliminated if the user profile has been stored for a period of time beyond a threshold level.
In another embodiment, a user profile is considered no longer useful if each interest score in the user profile is below a predefined threshold level. For example, where the interest score decays rapidly, the user profile may be eliminated unless the user visits the website very frequently. This mechanism can help the cleanup application 118 to ensure "that the user profile database 122 keeps user profiles of users genuinely interested in short-term content.
In process block 224, the reporting application 120 queries the user profile database 122 and reports the user profile information. In one embodiment, the reporting application 120 queries the user profile database 122 and utilizes the extracted information to forecast inventory, adjust ad and content delivery strategy, or retrieve any information on the user profiles.
In process block 220, the webpage requested by the user is transmitted from the web server application 110 to the web browser 102. The web server application 110 then retrieves the necessary data from the webpages database 112 in order to transmit the requested webpage. Figure 3 illustrates a flow diagram of a process 300 in which a user computer accesses the Internet website subsequent to the first time access. In process block 310, a webpage request and a cookie message are received from the user's computer 100. In one embodiment, because the webpage request occurs subsequent to a first webpage request, the user's computer already has an implanted cookie on the cookies database 104. When the user inputs the universal resource locator, the user's computer 100 automatically generates a webpage request with the message indicated by the cookie, namely the user identifier. When the web server application 110 receives the webpage request from the web browser 102, a cookie accompanies the webpage request which indicates to the web server application 110 that the user is a returning user.
Much like the first time a user requests the webpage, subsequent requests of the webpage triggers the above-mentioned of storing of webpage request information as shown in process block 216, populating the user profile database 218 as shown in process block 218, and performing cleanup and reporting operations as shown in process blocks 22 and 224, respectively. Therefore, as stated above, in process block 216 the user identifier, the request time, and the universal resource locator requested are stored in the log files database 124. In process block 218, the scoring application 126 can populate the user profile database 122 with a recalculated interest score. In addition, in process block 222 unused user profiles are eliminated, and in process block 224 user profiles can be outputted on a report format.
In process block 312, a user profile is retrieved from the user profile database 122 by the ad delivery application 114. In one embodiment, after the web server 110 receives the webpage request, the web server 110 sends the user identifier to the ad delivery application 114. The ad delivery application 114 utilizes the user identifier to query the user profile database 122 and obtain the user profile corresponding to the user requesting the webpage. With the user profile available, the ad delivery application 114 can generate an ad tailored to the interests of the user.
In process block 314, an ad is generated based on the user profile retrieved from the user profile database 122. In one embodiment, the ad delivery application 114 retrieves user profiles directly from the user profile database 122. In another embodiment, the user profile binary information can be decoded by a user segmentation application layer and later sent to the ad delivery application 114 in a format compatible with the ad delivery application 114.
As such, the segmentation application layer decodes the user profile and makes the user profile available to the ad delivery application 114. Furthermore, other applications located at the merchant server 130 can also retrieve user profiles efficiently without having to access and decode the user profile information from the user profile database 122.
In another embodiment, the user profile information is decoded dynamically. When the ad delivery application 114 receives a user identifier, the user identifier is used to query the segmentation application layer. If the user profile exists in the user profile database 122, the user profile can be decoded and modified to adjust the interest score that indicates the current interest level for each segment according to the current webpage request. A communication protocol between the ad delivery application 114 and the segmentation application layer is applied to transform the profile into a string of data that can be used directly by the ad delivery application 114 to apply the targeting logic.
Furthermore, the user profile database 122 can be updated immediately after receiving webpage request such that the user profile retrieved by the ad delivery application 114 includes the latest webpage request. For example, if the user requested a sports webpage, the webpage request information is reflected on the user profile, such that when ad delivery application 114 access the user profile, the user profile includes the webpage request for the sports webpage. Thus, the ad delivery application 114 can utilize interest scores that are updated with realtime data so the webpage requested includes behavioral ad targeting that takes into account the current webpage request. In one embodiment, the ad delivery application 122 can be further utilized for inventory. An inventory management system can be utilized in tandem with the ad delivery application 122 in order to forecast inventory needs. The interest scores retrieved from the user profile database 122 and the advertisement presented to the users can be computed to calculate the demand for merchant products.
In another embodiment, the ad delivery application 122 can be used for targeted content delivery. Interest scores can further be used by the web server application 110 to tailor content displayed to the user. By providing personalized content, the user overall experience is enhanced thus increasing the probabilities that the user can continue requesting webpages from the merchant.
Finally, as stated above, in process block 320 the webpage requested by the user is transmitted from the web server application 110 to the web browser
102. However, because the ad delivery application 122 has generated an ad for the user, the web server application 110 can include the generated ad as part of the requested webpage.
Webpages with similar content can be classified under similar content categories. As mentioned earlier, human behavior is considered as a consistent process, and therefore online users consistently access the same or similar websites. Grouping similar webpages into content categories facilitates the segmentation of user behavior because a user can be assigned an interest score for each content category. CONTENT CATEGORIES
Figure 4A illustrates content categories and web pages associated to the content categories. Universal resource locators and corresponding content categories can be stored in a local database in the merchant server 108 or in the logging server 124. In one embodiment, the local database includes a relational table of content categories and universal resource locators associated with at least one content category.
Each content category can reflect a common area of interest. For example, a website 402 for the Entertainment and Sports Programming Network (ESPN) can have a content category 404 for Golf, a content category 406 for Formula One
Car Racing, and a content category 408 for IndyCar Car Racing. One or more webpages can be associated with each content category.
For example, a webpage 416 related to the Golf US Open can be associated with the content category 404. When a user requests webpage 416, an identifier for webpage 416 is matched with the content category that had been previously associated with the webpage 416, namely content category 404. A user requesting a webpage 416 about the GoTf US Open indicates a general level of interest in the content category 404 corresponding to Golf. Likewise, a webpage 418 related to the Golf Players can be matched with the content category 404. When a user requests webpage 418, the request can be associated with content category 404. A user that is interested in information on Golf Players can be assumed to be interested, at least partially, in Golf.
In another example, a user that is interested in Formula One Drivers,
Formula One Cars, and Car Racing News can be assumed to be generally interested in Formula One Car Racing. Thus, a webpage 420 that includes content on Formula One Drivers can be associated with content category 406 for
Formula One Car Racing. A webpage 422 related to the Formula One Cars can be associated with the content category 406. A webpage 424 related to Car Racing News can also be associated with Formula One Car Racing. Then, if a user requests webpage 424, the request is matched with content category 406 reflecting the user interest in Formula One Car Racing. In yet another example, a content category 408 corresponding to IndyCar
Car Racing can have one or more webpages associated with the content category 408. A webpage 426 that includes content on IndyCar Drivers, and a webpage 428 that includes content on IndyCar Cars can be associated with content category 406 for IndyCar Car Racing. In one embodiment, a user that visits the car racing news webpage 424 can also be deemed to be interested in IndyCar Car Racing. As such, a webpage can be associated with one or more content categories. Webpage 424, for example, can be associated with both Formula One Car Racing and IndyCar Car Racing. Thus, when a user requests the car racing new website 424, the user can be assumed to be interested in Formula One Car Racing and IndyCar Car Racing.
As previously explained, whenever a user requests a webpage, the user is assumed to haveTa certain level of interest in~the~associated content category. In order to profile the interest of the user in individual content categories a user profile is developed for each user.
USER PROFILING
Figure 4B illustrates the data structure of a user profile. Each user profile 430 can be implemented utilizing known data structures such as tables, database records, instances of an object-oriented class (an object), etc. In one embodiment, a user profile 430 can be a record of a user profile table 432 located in database 122. In another embodiment, a user profile can be an instance of a class having multiple attributes. Each user profile 430 can be uniquely identified by a user identifier. For example, the user profile table 432 can include a plurality of user identifiers and the data for each corresponding user profile.
In one embodiment, each profile 430 can be a generic map that includes values corresponding to the content category identifier 440, the interest score 436, and the timestamp 438. The content category identifier 440 can be numerical value that is interpreted as a content category. In another embodiment, the content category identifier 440 can be character string corresponding to the name of the content category, for example if the content category is Golf, the category identifier 440 can be a character string "Golf." In another embodiment, the interest score 436 is a point system represented by a numerical value with a range of 0 to 100 points. In another embodiment, the interest score 436 is a numerical value with a range of 0.00 to 1.00 points. In yet another embodiment, the interest score 436 can have any other numerical value range or be stored as a character string. Furthermore, the timestamp 436 can also be stored as a numerical value as well as a character string.
While utilizing various data types for storing the category content identifier 440, the interest score 436, and the timestamp 438, a numerical data type can be used by any of these. Having the content category identifier 440, the interest score 436, and the timestamp 438 stored in numerical values facilitates quick access to information as well as the ability to store larger numbers of user profiles 430. In addition, user profiles can be stored for longer periods of time without having to purge them in order to save memory space. Furthermore, utilizing numerical values to represent user interest prevents the usage of private information of users. This protects the user's privacy without sacrificing the effectiveness of the behavioral targeting system. The user profile 430 includes illustrated in Figure 4B, includes three content categories: Golf, Formula One Car Racing, and IndyCar Car Racing. The interest score 436 for Golf is 25 points, and the last request was made at 03/06/2005 at 8:32 AM. The user profile 430 in Figure 4B indicates that the user has in interest level of 15 points for the content category Formula One Car Racing and that the last visit to a webpage associated with the Formula One Car Racing content category was on 8/21/2004 at 7:57 PM. Finally, the user category for IndyCar Car Racing shows that the user does not have an interest in this content category. In another embodiment, the user profile can simply not include content category in which the user has no interest. For example, a user profile 430 can simply have one entry for Golf as the content category identifier 440, the interest score 436 as 25 points and the timestamp 438 of the last request. In another example, the user profile can have five content categories, each with a corresponding interest score 436 and a timestamp 438.
SCORE CALCULATION
The interest score 436 corresponding to each of the content categories can be calculated according to different contributing factors. The factors that can determine the interest score can include 1) the frequency of visits to any of the webpages associated with the content category; 2) the time elapsed from one visit to the next; and 3) a decay rate.
In one embodiment, the interest score 436 changes based on the frequency of visits to any of the webpages associated with the content category. The more frequent a user visits any of the webpages associated with the content category, the higher the interest score 436. Thus, every time a user visits any of the webpages associated with the content category, the interest score 436 can be increased by a fixed amount. On every visit to a given webpage, the content category associated the webpage is identified. For example, if a user visits the Formula One Car Racing webpage 428 illustrated in Figure 4A, the system maps the universal resource locator associated with the Formula One Car Racing webpage 428 to the corresponding category Formula One Car Racing 406. Based on the user identifier, the user profile 430 is retrieved and the interest score 436 for Formula One Car Racing is updated. The interest score 436 can be increased by 1, 5, 10 or any other arbitrary setting configured in the system to be the standard increase for a visit to a webpage in the content category. As such, in one embodiment, the amount of increase can be configured to be, for example, 10 points every time the user visits a webpage in the content category. In another embodiment, the amount of increase can be 5 points.
As such, the interest score 436 for a content category can increase based on the total number of visits to the webpages associated with the content. For example, referring to Figure 4A, visits the webpage 416 about the Golf US Open as well as visits to webpage 418 related to the Golf Players increases the interest score 436 of the content category Golf in the user profile 430. In the user profile
430 illustrated in Figure 4B, for instance, the fixed increase can be configured to
-be 10 points. If the user visits both, the webpage 416 for Golf US Open as well as visits to webpage 418 related to the Golf Players, the total increase would be 20 points for the content category Golf, 10 for each webpage visit. The interest score 436 increases from 25 points to 45 points.
In another embodiment, if a webpage is associated with more than one content category, each of the interest scores in the content categories can be increased. For example, in Figure 4A the car racing news webpage 424 is associated with two content categories, namely content category 408 corresponding to IndyCar Car Racing, and the content category 408 corresponding to Formula One Car Racing. Thus, when a user requests the car racing new website 424, the interest score 436 corresponding to each of the content categories increases. In the user profile illustrated in Figure 4B, for instance, if the fixed increase is configured to be 10 points, the interest score 436 for Formula One Car Racing increases from 15 points to 25 points. Likewise, the interest score 436 for IndyCar Car Racing will now be 10 points.
In another embodiment, the interest score 436 changes based on the time elapsed from one visit to the next. While the number of visits over a certain period of time is directly proportional to the interest score 436, the time between visits can also provides with insightful information about user behavior. For example, a first user might visit the Formula One Car Racing webpage four times in a period of twelve weeks. The four visits can occurred within a period of four minutes during the first week. A second user might also visit Formula One Car Racing webpage four times in a period of twelve weeks. However, the second user might visit once a week for the last four weeks of the twelve-week period. While the visitation frequency of both users is the same, namely four visits in twelve weeks, the behavior is not the same. To better account for this disparity, the elapsed time between visits can also be taken into account when calculating the interest score 436 for each category. In that way, the user profile 430 corresponding to each user is a closer- prediction of user's- interest for certain content. -
In one embodiment, the elapsed time is calculated by subtracting the current time from the timestamp 438 the last visit. As discussed above, the user profile 430 include the timestamp 438 he last visit to any of the webpages associated with the content category 440. For example, if the content category is Golf, the timestamp 438 can be for the last visit to the webpage 416 related to the Golf US Open. If the user then visits the same webpage 416, two days later, the calculated time elapsed will be two days. In another example, if the timestamp 438 can be for the last visit to the webpage 416 related to the Golf US Open and the user visits another webpage 418 for Golf Players two days later, the elapsed time will also be two days because webpage 418 and webpage 416 correspond to the same content category Golf. A user that visits a webpage within short intervals of time is assumed to have higher interest in the content of the webpage, and therefore in the content category. As such, the relationship between the interest score and the amount of time elapsed is inversely proportional. The shorter the time elapsed from one visit to the next, the higher the interest score. In one embodiment, the elapsed time can be expressed in terms of a multiplier fraction that has the amount of time in the denominator and one in the numerator. The elapsed time can be expressed in number of minutes, hours, etc.
In yet another embodiment, the interest score 436 changes based on a decay rate. A previously mentioned, the interest score 436 is decayed based on a configured decay rate. Users that visit websites of a certain content might not come back to visit for a long time, if ever. The current system therefore applies a decay rate for every user interest score 436. Thus, if the user continues to visit the website the interest score 436 is maintained, but if the user ceases to visit websites of the given content category, the user interest score 436 decreases to the point that the system no longer recognizes the user as being interested in the
-content category. This decay mimics the human memory function of forgetting where repeated access to a memory maintains that memory, but non-access to the memory eliminates the memory over time. The decay rate applied to the interest score functions as a "forgetting" mechanism. If a user does not revisit a webpage for a content category, the user interest in the category decreases over time, which is reflected in the decreasing interest score 436.
In one embodiment, the decay rate is applied uniformly to all interest scores 436 of all the user profiles. As such, the system can decay interest scores on a regular basis. In one embodiment, all interest scores of all user profiles are decayed by a percentage (e.g. 7.5%) everyday. In one embodiment, all interest scores of all user profiles are decayed by a percentage (e.g. 15%) every week. In yet another embodiment, the decay is a constant fixed number. For example, if interest score range is 1 to 100 points, the decay rate can be 1 point per day. In yet another embodiment, each interest score is decayed every time the user visits certain category, wherein the decay takes into consideration the amount of time elapsed such that the higher the time elapsed, the higher the decay rate applied. In another embodiment, a long-term and short-term decay rates can be used. When a user visits webpages associated with a particular content category a high number of times, the user can be assumed to have a long-term interest in that content category. For example, some content categories can have seasonal information that by definition greatly attracts users for short period of time until the next season arrives. If the user makes many visits to webpages related to a specific content category, the user can be deemed to have a long-term interest in that content category. In order to preserve the user profile in the system, the decay rate applied can be a long-term decay rate that decays the interest score for the seasonal content category much slower. Figure 4C illustrates a user profile and associated short-term decay rates.
In one embodiment, each user profile in the system can be divided into two portions:" a firsTpoftiόn containing content categories and corresponding interest" scores that decay with a short-term decay rate, and a second portion containing content categories and corresponding interest scores that decay with a long-term decay rate. Depending upon the data structure utilized, the user profile 430 can be configured to differentiate between the content categories in which the user has a long-term interest and a short-term interest.
For example each of the content category in a the user profile 430 can have a corresponding attribute or flag (implemented with any data type) that can indicate whether the decay applied to the interest score that particular content category is to be decayed with the short-term decay rate or the long term decay rate. In another example, each user profile 430 can include two tables, one for the content categories with long-term interest, and another for the content categories with short-term interest.
In one embodiment, when a user first accesses a particular content, that is any webpage associated with a particular content, the user is assumed to have short-term interest. As shown in Figure 4C, the user profile 430 can have all of the content categories having a short term applied. Namely, the decay rate attributes
442, 444, and 446 are all "Short Term," showing that the user has a short-term interest in Golf, Formula One Car Racing and IndyCar Car Racing. As discussed above, the decay can be calculated periodically for all of the interest scores in the system. The three interest scores in user profile 430 as shown in Figure 4C can decrease gradually but quickly as the short-term decay applies.
The interest score 436 for each category can increase, decrease or remain approximately the same. In order to remain the same, the user must frequently visit webpages associated with the content category. As the decay is applied to the interest score of the category, the decrease due to the decay can be compensated by the increase due to visits to webpages corresponding to the content category.
In another embodiment, a low threshold can be established such that when the interest score decreases due to decay applied, interest score that reaches the low threshold indicates to the system that the user is no longer interested in the content category corresponding to the low interest score. The content category can then be removed from the content categories in the user profile. If the content category was the only one left in the listing of content categories, the user profile can be removed from the user profile database. In yet another embodiment, a high threshold can be established such that when the interest score increases due very frequent visitation to webpages associated with the content category. When the interest score reaches the high threshold, the system can recognize the user as a "power user" and change the decay rate applied to the content category to be a long-term decay.
Figure 4D illustrates a user profile and associated short-term and long-term decay rates. When a user visits webpages related to the content category with a frequency that causes the interest score to increase and reach the high threshold level, the user can be considered to have a long-term interest in that content category. For example, a threshold level for a category "Golf can be established to be 40 points. Any user that reaches 40 points or beyond qualifies as a "power user" on the category of Golf, and therefore the decay rate applied can be a long- term decay rate. If the user has an interest score of 25 points for the Golf content category, the user can visit a Golf-related webpage twice in one day and accumulate 20 extra points (assuming 10 points are awarded per visit). Taking into consideration that a daily short-term decay rate of 7.5% was applied earlier, and without taking into consideration the elapsed time between visits, the total interest score for the user can be (25 - 1.875) + 20 = 43.125. Thus, the interest score for the Golf category for that user has surpassed the threshold level. Once the interest score reaches the high threshold, the decay rate applied 442 changes from short-term to long-term. Subsequently, all further application of decay to the interest level of the section golf can be long-term. The decay rate attributes 444 and 446 remain as "Short Term," reflecting that the user has a short-term interest in Formula One Car Racing and IndyCar Car Racing. The decay rate attribute 442 is now "Long Term" reflecting that the user has a long-term interest in Golf.
In another example, the high threshold level can be 50 points. If the user visits webpages associated with the Golf content category four times in day one, the interest score can increase to 40 points. Then in day two, if the user does not revisit a webpage associated with the Golf content category the user score can decay to 37 points. In day three, however, if the user visits two webpages associated with the Golf content category the interest score for the Golf content category can be (37 - 2.775) + 20 = 54.225 thus surpassing the threshold level of 50 points.
As discussed above, the decay can be calculated periodically for all of the interest scores in the system. The system applies the long-term decay rate to the interest scores that are flagged to have a long-term decay, and a short-term decay rate to the interest score that are flagged to have a short-term decay. For example, if the short-term decay rate is 20% and the long-term decay rate is 7.5%, all interest scores in all user profiles that are associated with content categories that have a long-term decay rate applied will be decayed by 7.5%. Likewise, all interest scores in all user profiles that are associated with content categories that have a short-term decay rate applied are decayed by 20%.
In one embodiment, the high threshold level can be the same for all content categories (e.g. 50 points). In another embodiment, the high threshold level can be different for each content category (e.g. 50 points for Golf, and 40 points for Formula One Car Racing).
Figure 5A illustrates a decay curve wherein the decay rate is a short-term decay rate. The curve shows the decay of the interest score for a content category in which the decay rate applied is a 20% decay. The initial interest score is 40 points. Without further interaction by the user, such as user visits to webpages related to the content category, the interest score decreases rapidly. For example, point 402 shows that on the seventh day the interest score has decreased down to 10 points.
Figure 5B illustrates a decay curve wherein the decay rate is a long-term decay rate. The curve shows the decay of the interest score for a content category in which the decay rate applied is a 7.5% decay rate. The initial interest score is 40 points. Without further interaction by the user, such as user visits to webpages related to the content category, the interest score decreases but not as rapidly as with a decay rate of 20% as show in Figure 5A above. For example, point 404 of Figure 5B shows that not until the nineteenth day does the interest score decrease down to 10 points.
Although certain illustrative embodiments and methods have been disclosed herein, it will be apparent form the foregoing disclosure to those skilled in the art that variations and modifications of such embodiments and methods may be made without departing from the true spirit and scope of the art disclosed. For example, a user can show interest in a content category not only by viewing associated webpages, by also by accessing web content related to the content category. For example images, video, audio, text, ads, flash animations, and any other web content that is related to the substance of the particular content category.
Many other examples of the art disclosed exist, each differing from others in matters of detail only. Accordingly, it is intended that the art disclosed shall be limited only to the extent required by the appended claims and the rules and principles of applicable law.

Claims

WHAT is CLAIMED is:
1. A method of generating a user profile for advertisement over a computer network, comprising:
determining a webpage identifier that identifies a webpage;
defining a content category identifier that identifies a content category;
associating the webpage identifier with the content category identifier when web content depicted in the webpage is relevant to the content category;
receiving, at a web server, a request for the webpage, the request including the webpage identifier and a first user identifier, the request being received from a client computer connected to the computer network;
locating a user profile by determining whether the first user identifier matches a second user identifier corresponding to the user profile, wherein the user profile includes a user interest score corresponding to a user content category identifier; and
increasing the corresponding user interest score by a predetermined amount when the content category identifier associated with the requested webpage matches the user content category identifier, wherein the predetermined amount is established independently from the web content depicted in the webpage being viewed by the user.
2. The method of claim 1, wherein the webpage identifier is a universal resource locator.
3. The method of claim 1 , wherein the corresponding user interest score is indicative of a level of interest of the user in a content category.
4. The method of claim 1 , further comprising reducing the corresponding user interest score according to a decay rate.
5. The method of claim 4, wherein the corresponding user interest score is reduced daily regardless of whether a webpage request has been received at the web server.
6. The method of claim 4, wherein the corresponding user interest score is reduced periodically regardless of whether a webpage request has been received at the web server.
7. The method of claim 1 , further comprising reducing the corresponding user interest score according to a long-term decay rate which slowly reduces the corresponding user interest score.
8. The method of claim 7, wherein the corresponding user interest score is reduced by a long-term decay rate when the corresponding user interest score surpasses a high threshold.
9. The method of claim 1, further comprising reducing the corresponding user interest score according to a short-term decay rate which rapidly reduces the corresponding user interest score.
10. The method of claim 9, wherein the corresponding user interest score is reduced by a short-term decay rate when the corresponding user interest score decreases below a low threshold.
11. The method of claim 1 , further comprising reducing the corresponding user interest score according to a short-term decay rate which rapidly reduces the corresponding user interest score, wherein when the corresponding user interest score surpasses a high threshold the corresponding user interest score ceases to be reduced by the short-term decay and instead is reduced by a long-term decay rate.
12. The method of claim 1 , further comprising utilizing a plurality of decay rates to decay the corresponding user interest score, wherein a decay rate from the plurality of decay rates is chosen to be utilized depending upon the user behavior.
13. The method of claim 1 , further comprising decaying the corresponding user interest score of a first content category in the user profile at a different rate from the corresponding user interest score of a second content category, wherein the difference is due to the user behaving differently with respect to the first content category in comparison to the second content category.
14. The method of claim 13, wherein the corresponding user interest score of the first content category in the user profile is decayed slowly if the user shows a long-term interest in the content category.
15. The method of claim 14, wherein the user shows a long-term interest if the user makes webpages requests frequent enough such that the corresponding user interest score increases beyond a predefined threshold.
16. The method of claim 13, wherein the corresponding user interest score of the second content category in the user profile is decayed quickly if the user shows a short-term interest in the content category.
17. The method of claim 1 , further comprising: assigning the corresponding user interest score to be reduced by a long-term decay rate if the corresponding user interest score of the content category identifier surpasses a high threshold;
determining whether the user has a long-term interest in the content category by checking whether the content category is assigned to be reduced by a long-term decay rate; and
if it is determined that the user has a long-term interest in the content category, decaying the corresponding user interest score utilizing the long-term decay rate, wherein the long-term decay rate slowly reduces the corresponding user interest score.
18. The method of claim 17, wherein if it is determined that the user does not have a long-term interest in the content category, decaying the corresponding user interest score utilizing the short-term decay rate.
19. The method of claim 1, further comprising deleting the user profile if a webpage request is not received-from a user after a threshold time. -
20. The method of claim 1 , further comprising deleting the user profile if the corresponding user interest score decreases beyond a low threshold.
21. The method of claim 1 , further comprising deleting a user interest score corresponding to a particular content category from the user profile if the user interest score corresponding to that particular content category decreases beyond a low threshold.
22. The method of claim 1 , wherein the webpage request includes a webpage request timestamp.
23. The method of claim 1 , wherein the corresponding user interest score can be further increased by taking into account an elapsed time between the time of a last webpage request and the time of the webpage request received.
24. The method of claim 23, wherein the corresponding user interest score is increased in inversely proportion to the length of time of the elapsed time.
25. The method of claim 1 , wherein the predetermined amount is utilized to increase the corresponding user interest score every time the user makes a request to a website that has an associated content category.
26. The method of claim 1 , wherein the predetermined amount is utilized to increase the user interest score corresponding to each one of a plurality of content category identifiers in a user profile.
27. The method of claim 1 , further comprising logging the webpage identifier, and the user identifier, in a log file.
28. The method of claim 1 , further associating the webpage and at least one other content category by correlating the webpage identifier corresponding to the webpage with a content category identifier corresponding to the at least one other content category.
29. The method of claim 1 , wherein the webpage identifier is further associated with a second content category, such that the when the webpage is requested, a second user interest score corresponding to the second content category is increased by the predetermined amount.
30. The method of claim 1, wherein the user profile is implemented utilizing a data structure comprising a content category field, an interest score field, and a timestamp field.
31. A method of generating a user profile for advertisement over a computer network, comprising:
determining a webpage identifier that identifies a webpage;
defining a content category identifier that identifies a content category;
associating the webpage identifier with the content category identifier when web content depicted in the webpage is relevant to the content category;
receiving, at a web server, a request for the webpage, the request including the webpage identifier and a first user identifier, the request being received from a client computer connected to the computer network;
locating a user profile by determining whether the first user identifier matches a second user identifier corresponding to the user profile, wherein the user profile includes at least one user content category identifier and a corresponding user interest score;
providing a predetermined amount," wherein the predetermined amount is established independently from the web content depicted in the webpage being viewed by the user;
if the content category identifier matches the at least one user content category identifier, increasing the corresponding user interest score by the predetermined amount;
assigning the corresponding user interest score to be reduced by a long-term decay rate when the corresponding user interest score of the content category identifier surpasses a high threshold; determining whether the user has a long-term interest in the content category by checking whether the content category is assigned to be reduced by a long-term decay rate; and
if it is determined that the user has a long-term interest in the content category, decaying the corresponding user interest score utilizing the long-term decay rate, wherein the long-term decay rate slowly reduces the corresponding user interest score.
32. The method of claim 31 , wherein if it is determined that the user does not have a long-term interest in the content category, decaying the corresponding user interest score utilizing a short-term decay rate.
33. A system to deliver advertisement over a computer network, comprising:
a merchant sever configured to receive a webpage request from a web browser on the user computer, the webpage request including a universal resource locator and a user identifier; and
a user profile database to store user profiles, each user profile comprising a plurality of content category identifiers, wherein each of the plurality of content category identifier has a corresponding user interest score and a timestamp indicating when the webpage request is last received, wherein the corresponding user interest score increases based on the frequency a webpage associated with the content category is requested.
34. The system of claim 33, wherein the corresponding user interest score is indicative of a level of interest of the user in a content category identified by the at least one user content category identifier.
35. The system of claim 33, wherein the corresponding user interest score is reduced according to a decay rate.
36. The system of claim 35, wherein the corresponding user interest score is reduced daily regardless of whether a webpage request has been received at the web server.
37. The system of claim 35, wherein the corresponding user interest score is reduced according to a long-term decay rate which slowly reduces the corresponding user interest score.
38. The system of claim 37, wherein the corresponding user interest score is reduced by a long-term decay rate if the corresponding user interest score surpasses a high threshold.
39. The system of claim 33, wherein the corresponding user interest score is reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score.
40. The system of claim 39, wherein the corresponding user interest score is reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold.
41. A method of receiving a targeted advertisement over a computer network, comprising:
transmitting from a client computer a request for a webpage, the request including a webpage identifier and a first user identifier, the request being received by a web server that locates a user profile by determining whether the first user identifier matches a second user identifier corresponding to the user profile, wherein the web server further locates a content category identifier in the user profile by determining whether web content depicted in the webpage is associated to the content identified by the content category identifier, wherein the web server increases a corresponding user interest score by a predetermined amount that is independent from the web content depicted in the webpage;
receiving the requested webpage from the web server; and
receiving Internet content, wherein the Internet content is related to a content category in the user profile having an associated interest score higher than a predetermined amount.
42. The method of claim 41 , wherein the Internet content includes advertisement.
43. The method of claim 41 , wherein the Internet content includes music.
44. The method of claim 41 , wherein the Internet content includes video.
45. The method of claim 41 , wherein the webpage identifier is a universal resource locator.
46. The method of claim 41 , wherein the corresponding user interest score is indicative of a level of interest of the user in a content category identified by the at least one user content category identifier.
47. The method of claim 41 , wherein the corresponding user interest score is reduced according to a decay rate.
48. The method of claim 37, wherein the corresponding user interest score is reduced daily regardless of whether a webpage request has been received at the web server.
49. The method of claim 41 , wherein the corresponding user interest score is reduced according to a long-term decay rate which slowly reduces the corresponding user interest score.
50. The method of claim 49, wherein the corresponding user interest score is reduced by a long-term decay rate if the corresponding user interest score surpasses a high threshold.
51. The method of claim 41 , wherein the corresponding user interest score is reduced according to a short-term decay rate which rapidly reduces the corresponding user interest score.
52. The method of claim 51 , wherein the corresponding user interest score is reduced by a short-term decay rate if the corresponding user interest score decreases below a low threshold.
PCT/US2006/037482 2005-10-28 2006-09-26 System and method for targeted ad delivery WO2007055812A2 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US11/262,240 2005-10-28
US11/262,240 US7734632B2 (en) 2005-10-28 2005-10-28 System and method for targeted ad delivery

Publications (2)

Publication Number Publication Date
WO2007055812A2 true WO2007055812A2 (en) 2007-05-18
WO2007055812A3 WO2007055812A3 (en) 2009-04-23

Family

ID=37997766

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2006/037482 WO2007055812A2 (en) 2005-10-28 2006-09-26 System and method for targeted ad delivery

Country Status (2)

Country Link
US (2) US7734632B2 (en)
WO (1) WO2007055812A2 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012024066A2 (en) * 2010-08-19 2012-02-23 Yahoo! Inc. Contextual advertising with user features

Families Citing this family (157)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7124129B2 (en) * 1998-03-03 2006-10-17 A9.Com, Inc. Identifying the items most relevant to a current query based on items selected in connection with similar queries
ATE243869T1 (en) 1998-03-03 2003-07-15 Amazon Com Inc IDENTIFICATION OF THE MOST RELEVANT ANSWERS TO A CURRENT SEARCH QUERY BASED ON ANSWERS ALREADY SELECTED FOR SIMILAR QUERIES
US9819561B2 (en) 2000-10-26 2017-11-14 Liveperson, Inc. System and methods for facilitating object assignments
US8868448B2 (en) 2000-10-26 2014-10-21 Liveperson, Inc. Systems and methods to facilitate selling of products and services
US8620915B1 (en) 2007-03-13 2013-12-31 Google Inc. Systems and methods for promoting personalized search results based on personal information
RU2007123612A (en) * 2004-11-23 2008-12-27 Конинклейке Филипс Электроникс Н.В. (Nl) FILE MANAGEMENT METHOD AND DEVICE
US8874570B1 (en) 2004-11-30 2014-10-28 Google Inc. Search boost vector based on co-visitation information
US9065727B1 (en) 2012-08-31 2015-06-23 Google Inc. Device identifier similarity models derived from online event signals
US8738732B2 (en) 2005-09-14 2014-05-27 Liveperson, Inc. System and method for performing follow up based on user interactions
US9432468B2 (en) * 2005-09-14 2016-08-30 Liveperson, Inc. System and method for design and dynamic generation of a web page
US8171128B2 (en) * 2006-08-11 2012-05-01 Facebook, Inc. Communicating a newsfeed of media content based on a member's interactions in a social network environment
CN1987916A (en) * 2005-12-21 2007-06-27 腾讯科技(深圳)有限公司 Method and device for releasing network advertisements
GB2435114A (en) * 2006-02-08 2007-08-15 Rapid Mobile Media Ltd Providing targeted additional content
US7996396B2 (en) * 2006-03-28 2011-08-09 A9.Com, Inc. Identifying the items most relevant to a current query based on user activity with respect to the results of similar queries
US8438170B2 (en) * 2006-03-29 2013-05-07 Yahoo! Inc. Behavioral targeting system that generates user profiles for target objectives
US7814109B2 (en) * 2006-03-29 2010-10-12 Yahoo! Inc. Automatic categorization of network events
US8504575B2 (en) * 2006-03-29 2013-08-06 Yahoo! Inc. Behavioral targeting system
US20070239517A1 (en) * 2006-03-29 2007-10-11 Chung Christina Y Generating a degree of interest in user profile scores in a behavioral targeting system
US7809740B2 (en) * 2006-03-29 2010-10-05 Yahoo! Inc. Model for generating user profiles in a behavioral targeting system
US7904448B2 (en) * 2006-03-29 2011-03-08 Yahoo! Inc. Incremental update of long-term and short-term user profile scores in a behavioral targeting system
WO2007147080A1 (en) 2006-06-16 2007-12-21 Almondnet, Inc. Media properties selection method and system based on expected profit from profile-based ad delivery
US20080005282A1 (en) * 2006-07-03 2008-01-03 Dewey Gaedcke Method for displaying user generated content in a web browser
US20080005125A1 (en) * 2006-07-03 2008-01-03 Dewey Gaedeke System for displaying user generated content in a web browser
US20100169363A1 (en) * 2006-07-03 2010-07-01 Dewey Gaedcke System and method for managing and displaying additive content in a web browser
US7849079B2 (en) * 2006-07-31 2010-12-07 Microsoft Corporation Temporal ranking of search results
US7685199B2 (en) * 2006-07-31 2010-03-23 Microsoft Corporation Presenting information related to topics extracted from event classes
US7577718B2 (en) * 2006-07-31 2009-08-18 Microsoft Corporation Adaptive dissemination of personalized and contextually relevant information
GB2435565B (en) 2006-08-09 2008-02-20 Cvon Services Oy Messaging system
WO2008049955A1 (en) 2006-10-27 2008-05-02 Cvon Innovations Ltd Method and device for managing subscriber connection
EP2082366A2 (en) * 2006-11-01 2009-07-29 CVON Innovations Limited Optimization of advertising campaigns on mobile networks
GB2435730B (en) 2006-11-02 2008-02-20 Cvon Innovations Ltd Interactive communications system
GB2436412A (en) 2006-11-27 2007-09-26 Cvon Innovations Ltd Authentication of network usage for use with message modifying apparatus
US8140566B2 (en) * 2006-12-12 2012-03-20 Yahoo! Inc. Open framework for integrating, associating, and interacting with content objects including automatic feed creation
US20080140591A1 (en) * 2006-12-12 2008-06-12 Yahoo! Inc. System and method for matching objects belonging to hierarchies
US20090234814A1 (en) * 2006-12-12 2009-09-17 Marco Boerries Configuring a search engine results page with environment-specific information
US20090240564A1 (en) * 2006-12-12 2009-09-24 Marco Boerries Open framework for integrating, associating, and interacting with content objects including advertisement and content personalization
GB2440990B (en) 2007-01-09 2008-08-06 Cvon Innovations Ltd Message scheduling system
US20080201643A1 (en) 2007-02-01 2008-08-21 7 Billion People, Inc. System for creating customized web content based on user behavioral portraits
US8671016B2 (en) * 2007-02-28 2014-03-11 Ebay, Inc. Methods and apparatus for advertising via computer networks and websites
GB2445630B (en) 2007-03-12 2008-11-12 Cvon Innovations Ltd Dynamic message allocation system and method
US7860848B2 (en) * 2007-03-30 2010-12-28 Sap Ag Partial sub-lock for delta object management
KR20090000829A (en) * 2007-04-06 2009-01-08 엔에이치엔(주) Online advertising method for reflecting update of database in real time and system thereof
EP1990762A1 (en) * 2007-05-07 2008-11-12 Alcatel Lucent A system and associated method for selecting advertisements
US8572203B1 (en) * 2007-05-10 2013-10-29 United Services Automobile Association (Usaa) Systems and methods for personalized customer messages
GB2440408B (en) * 2007-05-16 2008-06-25 Cvon Innovations Ltd Method and system for scheduling of messages
US8935718B2 (en) 2007-05-22 2015-01-13 Apple Inc. Advertising management method and system
GB2450144A (en) 2007-06-14 2008-12-17 Cvon Innovations Ltd System for managing the delivery of messages
GB2452789A (en) 2007-09-05 2009-03-18 Cvon Innovations Ltd Selecting information content for transmission by identifying a keyword in a previous message
US8190475B1 (en) * 2007-09-05 2012-05-29 Google Inc. Visitor profile modeling
US20090099932A1 (en) * 2007-10-11 2009-04-16 Cvon Innovations Ltd. System and method for searching network users
GB2453810A (en) 2007-10-15 2009-04-22 Cvon Innovations Ltd System, Method and Computer Program for Modifying Communications by Insertion of a Targeted Media Content or Advertisement
US20090132368A1 (en) * 2007-10-19 2009-05-21 Paul Cotter Systems and Methods for Providing Personalized Advertisement
US8839088B1 (en) 2007-11-02 2014-09-16 Google Inc. Determining an aspect value, such as for estimating a characteristic of online entity
US20090119172A1 (en) * 2007-11-02 2009-05-07 Soloff David L Advertising Futures Marketplace Methods and Systems
US20090132365A1 (en) * 2007-11-15 2009-05-21 Microsoft Corporation Search, advertising and social networking applications and services
US20090164287A1 (en) * 2007-12-24 2009-06-25 Kies Jonathan K Method and apparatus for optimizing presentation of media content on a wireless device based on user behavior
US8751492B1 (en) * 2008-01-17 2014-06-10 Amdocs Software Systems Limited System, method, and computer program product for selecting an event category based on a category score for use in providing content
US20090187593A1 (en) * 2008-01-17 2009-07-23 Qualcomm Incorporated Methods and Apparatus for Targeted Media Content Delivery and Acquisition in a Wireless Communication Network
US8768922B2 (en) * 2008-02-08 2014-07-01 Microsoft Corporation Ad retrieval for user search on social network sites
US20090307002A1 (en) * 2008-02-13 2009-12-10 Marketing Technology Solutions System and Method for Communicating Targeted Health Related Data
US20090258637A1 (en) * 2008-04-11 2009-10-15 Beijing Focus Wireless Media Technology Co., ltd. Method for user identity tracking
US20090300009A1 (en) * 2008-05-30 2009-12-03 Netseer, Inc. Behavioral Targeting For Tracking, Aggregating, And Predicting Online Behavior
US8078974B2 (en) * 2008-06-27 2011-12-13 Microsoft Corporation Relating web page change with revisitation patterns
US8260846B2 (en) 2008-07-25 2012-09-04 Liveperson, Inc. Method and system for providing targeted content to a surfer
US8762313B2 (en) 2008-07-25 2014-06-24 Liveperson, Inc. Method and system for creating a predictive model for targeting web-page to a surfer
JP5327784B2 (en) * 2008-07-30 2013-10-30 株式会社日立製作所 Computer system, information collection support device, and information collection support method
US8805844B2 (en) 2008-08-04 2014-08-12 Liveperson, Inc. Expert search
US20100042496A1 (en) 2008-08-13 2010-02-18 Disney Enterprises, Inc. Advertising inventory management system and method
US20100042485A1 (en) * 2008-08-13 2010-02-18 Disney Enterprises, Inc. Advertisement delivery system
US9892417B2 (en) 2008-10-29 2018-02-13 Liveperson, Inc. System and method for applying tracing tools for network locations
US8694492B2 (en) * 2008-11-03 2014-04-08 Telefonaktiebolaget Lm Ericsson (Publ) Pre-fetching of data in a mobile communications environment
EP2239700A1 (en) * 2009-02-13 2010-10-13 Rapid Mobile Media Ltd. A system and method for determining user response to wireless messages
US20100223105A1 (en) * 2009-02-27 2010-09-02 Kindsight, Inc. Method and apparatus for content targeting one user group based on behavioral profiling of another user group
US8566332B2 (en) * 2009-03-02 2013-10-22 Hewlett-Packard Development Company, L.P. Populating variable content slots on web pages
US8352319B2 (en) * 2009-03-10 2013-01-08 Google Inc. Generating user profiles
US20120046995A1 (en) * 2009-04-29 2012-02-23 Waldeck Technology, Llc Anonymous crowd comparison
US8326999B2 (en) * 2009-06-02 2012-12-04 International Business Machines Corporation Creating context-sensitive webpage time-out intervals
US20110022465A1 (en) * 2009-07-24 2011-01-27 Prasannakumar Jobigenahally Malleshaiah System and Method for Managing Online Experiences Based on User Sentiment Characteristics and Publisher Targeting Goals
US9213767B2 (en) * 2009-08-10 2015-12-15 Hewlett-Packard Development Company, L.P. Method and system for characterizing web content
US8819172B2 (en) 2010-11-04 2014-08-26 Digimarc Corporation Smartphone-based methods and systems
US20110138401A1 (en) * 2009-12-04 2011-06-09 Microsoft Corporation Live update of user segments
US8244760B2 (en) 2009-12-04 2012-08-14 Microsoft Corporation Segmentation and profiling of users
CN102656570B (en) * 2009-12-16 2015-11-25 Idtp控股公司 For method and the server of buffer memory
CN102227892B (en) * 2009-12-24 2014-01-01 华为终端有限公司 Method, device and system for delivering advertisement
US8204878B2 (en) * 2010-01-15 2012-06-19 Yahoo! Inc. System and method for finding unexpected, but relevant content in an information retrieval system
US20110208585A1 (en) * 2010-02-19 2011-08-25 Peter Daboll Systems and Methods for Measurement of Engagement
EP2556449A1 (en) 2010-04-07 2013-02-13 Liveperson Inc. System and method for dynamically enabling customized web content and applications
US9990641B2 (en) 2010-04-23 2018-06-05 Excalibur Ip, Llc Finding predictive cross-category search queries for behavioral targeting
US8898217B2 (en) 2010-05-06 2014-11-25 Apple Inc. Content delivery based on user terminal events
US8370330B2 (en) 2010-05-28 2013-02-05 Apple Inc. Predicting content and context performance based on performance history of users
US8504419B2 (en) 2010-05-28 2013-08-06 Apple Inc. Network-based targeted content delivery based on queue adjustment factors calculated using the weighted combination of overall rank, context, and covariance scores for an invitational content item
US20130066708A1 (en) * 2010-06-04 2013-03-14 Ad Dynamo International (Pty) Ltd. Online advertising system and a method of operating the same
US8510658B2 (en) 2010-08-11 2013-08-13 Apple Inc. Population segmentation
US8640032B2 (en) 2010-08-31 2014-01-28 Apple Inc. Selection and delivery of invitational content based on prediction of user intent
US8983978B2 (en) 2010-08-31 2015-03-17 Apple Inc. Location-intention context for content delivery
US8510309B2 (en) 2010-08-31 2013-08-13 Apple Inc. Selection and delivery of invitational content based on prediction of user interest
US8843817B2 (en) * 2010-09-14 2014-09-23 Yahoo! Inc. System and method for obtaining user information
US9679060B2 (en) * 2010-10-13 2017-06-13 Microsoft Technology Licensing, Llc Following online social behavior to enhance search experience
KR101578802B1 (en) * 2010-11-17 2015-12-21 네이버 주식회사 System and method for providing advertisement based on user's interest continuity on keyword
US8918465B2 (en) 2010-12-14 2014-12-23 Liveperson, Inc. Authentication of service requests initiated from a social networking site
US9350598B2 (en) 2010-12-14 2016-05-24 Liveperson, Inc. Authentication of service requests using a communications initiation feature
US8688524B1 (en) 2011-06-28 2014-04-01 Amazon Technologies, Inc. Tracking online impressions to offline purchases
US9785968B1 (en) 2011-07-01 2017-10-10 Google Inc. Selecting content based on user actions and preferences associates with a same time period in a previous year
US9098600B2 (en) * 2011-09-14 2015-08-04 International Business Machines Corporation Deriving dynamic consumer defined product attributes from input queries
US8903909B1 (en) * 2011-09-15 2014-12-02 Google Inc. Detecting and extending engagement with stream content
WO2013116974A1 (en) 2012-02-06 2013-08-15 Empire Technology Development Llc Web tracking protection
US8943002B2 (en) 2012-02-10 2015-01-27 Liveperson, Inc. Analytics driven engagement
US8805941B2 (en) 2012-03-06 2014-08-12 Liveperson, Inc. Occasionally-connected computing interface
US9930093B2 (en) * 2012-03-14 2018-03-27 International Business Machines Corporation Dynamic web session clean-up
IL219244A0 (en) * 2012-04-18 2012-07-31 Google Inc Splitting of user-lists
US9563336B2 (en) 2012-04-26 2017-02-07 Liveperson, Inc. Dynamic user interface customization
US9053185B1 (en) 2012-04-30 2015-06-09 Google Inc. Generating a representative model for a plurality of models identified by similar feature data
US8527526B1 (en) 2012-05-02 2013-09-03 Google Inc. Selecting a list of network user identifiers based on long-term and short-term history data
US20130305272A1 (en) * 2012-05-14 2013-11-14 Tellabs Operations, Inc. Procedure, apparatus, system, and computer program for video addressable advertising
US9672196B2 (en) 2012-05-15 2017-06-06 Liveperson, Inc. Methods and systems for presenting specialized content using campaign metrics
US8914500B1 (en) 2012-05-21 2014-12-16 Google Inc. Creating a classifier model to determine whether a network user should be added to a list
US8886575B1 (en) 2012-06-27 2014-11-11 Google Inc. Selecting an algorithm for identifying similar user identifiers based on predicted click-through-rate
US9141504B2 (en) 2012-06-28 2015-09-22 Apple Inc. Presenting status data received from multiple devices
US8874589B1 (en) 2012-07-16 2014-10-28 Google Inc. Adjust similar users identification based on performance feedback
US8782197B1 (en) 2012-07-17 2014-07-15 Google, Inc. Determining a model refresh rate
US8856110B2 (en) * 2012-08-01 2014-10-07 Meterfy Ltd. Method and apparatus for providing a response to a query
CN103577478B (en) * 2012-08-06 2015-07-29 腾讯科技(深圳)有限公司 Web page push method and system
US8849804B1 (en) * 2012-08-10 2014-09-30 Google Inc. Distributing interest categories within a hierarchical classification
US20150235215A1 (en) * 2012-08-16 2015-08-20 Tango Mobile, LLC System and Method for Mobile or Web-Based Payment/Credential Process
US9400871B1 (en) * 2012-08-27 2016-07-26 Google Inc. Selecting content for devices specific to a particular user
US8886799B1 (en) 2012-08-29 2014-11-11 Google Inc. Identifying a similar user identifier
US8910282B2 (en) * 2012-09-04 2014-12-09 Jonathan Somers System and method for protecting devices on dynamically configured network
US8745074B1 (en) * 2012-09-14 2014-06-03 Google Inc. Method and system for evaluating content via a computer network
US20140046804A1 (en) * 2012-10-22 2014-02-13 Mojo Motors, Inc. Customizing online automotive vehicle searches
US20140156627A1 (en) * 2012-11-30 2014-06-05 Microsoft Corporation Mapping of topic summaries to search results
US9009480B1 (en) 2013-03-07 2015-04-14 Facebook, Inc. Techniques for handshake-free encrypted communication using public key bootstrapping
US20140258890A1 (en) * 2013-03-08 2014-09-11 Yahoo! Inc. Systems and methods for altering the speed of content movement based on user interest
US9549047B1 (en) * 2013-03-14 2017-01-17 Google Inc. Initiating a client-side user model
EP2973245A4 (en) 2013-03-15 2017-01-11 Factual Inc. Crowdsouorcing domain specific intelligence
US9906608B2 (en) * 2013-04-30 2018-02-27 International Business Machines Corporation Intelligent adaptation of mobile applications based on constraints and contexts
US9866648B2 (en) * 2013-05-10 2018-01-09 Laurent Bortolamiol Automatic transmission of user profile information to a web server
US9659065B1 (en) * 2013-06-05 2017-05-23 Google Inc. Ranking search results based on current or past presences
US9767489B1 (en) * 2013-08-30 2017-09-19 Google Inc. Content item impression effect decay
US20150149305A1 (en) * 2013-11-26 2015-05-28 Jack Ke Zhang Triggered presentation of third-party interactive content channels on electronic devices
US9311639B2 (en) 2014-02-11 2016-04-12 Digimarc Corporation Methods, apparatus and arrangements for device to device communication
US9959317B2 (en) 2014-03-05 2018-05-01 Empire Technology Development Llc Semantic data generation
US20150278353A1 (en) * 2014-03-31 2015-10-01 Linkedln Corporation Methods and systems for surfacing content items based on impression discounting
US11386442B2 (en) 2014-03-31 2022-07-12 Liveperson, Inc. Online behavioral predictor
US10031976B2 (en) * 2014-04-07 2018-07-24 Paypal, Inc. Personalization platform
US9348493B2 (en) 2014-05-13 2016-05-24 Jack Ke Zhang Automated subscriber-based customization of electronic channels for content presentation
US20170337589A1 (en) * 2014-10-27 2017-11-23 Yintao Yu Cross device identity matching for online advertising
US11093950B2 (en) * 2015-02-02 2021-08-17 Opower, Inc. Customer activity score
US10142908B2 (en) 2015-06-02 2018-11-27 Liveperson, Inc. Dynamic communication routing based on consistency weighting and routing rules
US20170357987A1 (en) * 2015-06-09 2017-12-14 Clickagy, LLC Online platform for predicting consumer interest level
US10783534B2 (en) * 2015-06-09 2020-09-22 Clickagy, LLC Method, system and computer readable medium for creating a profile of a user based on user behavior
US20170103073A1 (en) * 2015-10-08 2017-04-13 Quixey, Inc. Identifying Expert Reviewers
US10311037B2 (en) 2015-10-08 2019-06-04 Signal Vine, Llc Systems and methods for providing a two-way, intelligent text messaging platform
US10445776B2 (en) 2015-11-10 2019-10-15 International Business Machines Corporation User-configurable settings for web-based advertising
US11423443B2 (en) * 2016-02-05 2022-08-23 Fredrick T Howard Time limited media sharing
WO2017136363A1 (en) * 2016-02-05 2017-08-10 Howard Fredrick T Time limited image sharing
CN116528220A (en) 2016-08-14 2023-08-01 利维帕尔森有限公司 Method and system for real-time remote control of mobile applications
US20180308016A1 (en) * 2017-04-13 2018-10-25 Tripaly, LLC Method for allocating logistical travel resources

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6415322B1 (en) * 1998-02-27 2002-07-02 Engage, Inc. Dual/blind identification
US6421675B1 (en) * 1998-03-16 2002-07-16 S. L. I. Systems, Inc. Search engine
US20020111994A1 (en) * 2001-02-14 2002-08-15 International Business Machines Corporation Information provision over a network based on a user's profile
US20030128236A1 (en) * 2002-01-10 2003-07-10 Chen Meng Chang Method and system for a self-adaptive personal view agent
US20070061195A1 (en) * 2005-09-13 2007-03-15 Yahoo! Inc. Framework for selecting and delivering advertisements over a network based on combined short-term and long-term user behavioral interests

Family Cites Families (50)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6202058B1 (en) 1994-04-25 2001-03-13 Apple Computer, Inc. System for ranking the relevance of information objects accessed by computer users
US5636346A (en) * 1994-05-09 1997-06-03 The Electronic Address, Inc. Method and system for selectively targeting advertisements and programming
US6029195A (en) 1994-11-29 2000-02-22 Herz; Frederick S. M. System for customized electronic identification of desirable objects
US6460036B1 (en) 1994-11-29 2002-10-01 Pinpoint Incorporated System and method for providing customized electronic newspapers and target advertisements
GB9522791D0 (en) * 1995-11-07 1996-01-10 Cambridge Consultants Information retrieval and display systems
US5848396A (en) * 1996-04-26 1998-12-08 Freedom Of Information, Inc. Method and apparatus for determining behavioral profile of a computer user
US6014638A (en) 1996-05-29 2000-01-11 America Online, Inc. System for customizing computer displays in accordance with user preferences
US5727129A (en) * 1996-06-04 1998-03-10 International Business Machines Corporation Network system for profiling and actively facilitating user activities
US6092105A (en) 1996-07-12 2000-07-18 Intraware, Inc. System and method for vending retail software and other sets of information to end users
US5933827A (en) 1996-09-25 1999-08-03 International Business Machines Corporation System for identifying new web pages of interest to a user
US6038561A (en) 1996-10-15 2000-03-14 Manning & Napier Information Services Management and analysis of document information text
BR9714949A (en) 1996-12-20 2005-04-12 Princeton Video Image Inc Superior adjustment device for targeted electronic insertion of video indications
US6038560A (en) * 1997-05-21 2000-03-14 Oracle Corporation Concept knowledge base search and retrieval system
US6484149B1 (en) 1997-10-10 2002-11-19 Microsoft Corporation Systems and methods for viewing product information, and methods for generating web pages
US6009410A (en) 1997-10-16 1999-12-28 At&T Corporation Method and system for presenting customized advertising to a user on the world wide web
US6055510A (en) 1997-10-24 2000-04-25 At&T Corp. Method for performing targeted marketing over a large computer network
US5943670A (en) 1997-11-21 1999-08-24 International Business Machines Corporation System and method for categorizing objects in combined categories
US6029176A (en) 1997-11-25 2000-02-22 Cannon Holdings, L.L.C. Manipulating and analyzing data using a computer system having a database mining engine resides in memory
US6108645A (en) 1997-11-26 2000-08-22 International Business Machines Corporation Method and apparatus for efficient profile matching in a large scale webcasting system
DE19804192A1 (en) 1998-02-03 1999-08-12 Siemens Ag Method for producing a power semiconductor component
US6286005B1 (en) 1998-03-11 2001-09-04 Cannon Holdings, L.L.C. Method and apparatus for analyzing data and advertising optimization
US6327574B1 (en) 1998-07-07 2001-12-04 Encirq Corporation Hierarchical models of consumer attributes for targeting content in a privacy-preserving manner
AU5234999A (en) 1998-08-03 2000-02-28 Doubleclick Inc. Network for distribution of re-targeted advertising
US6266668B1 (en) 1998-08-04 2001-07-24 Dryken Technologies, Inc. System and method for dynamic data-mining and on-line communication of customized information
AU5465099A (en) 1998-08-04 2000-02-28 Rulespace, Inc. Method and system for deriving computer users' personal interests
US6154783A (en) 1998-09-18 2000-11-28 Tacit Knowledge Systems Method and apparatus for addressing an electronic document for transmission over a network
US6317722B1 (en) 1998-09-18 2001-11-13 Amazon.Com, Inc. Use of electronic shopping carts to generate personal recommendations
US6385619B1 (en) 1999-01-08 2002-05-07 International Business Machines Corporation Automatic user interest profile generation from structured document access information
JP2000330856A (en) 1999-05-21 2000-11-30 Nec Corp Information collection device and method therefor
US6578025B1 (en) 1999-06-11 2003-06-10 Abuzz Technologies, Inc. Method and apparatus for distributing information to users
US7072888B1 (en) * 1999-06-16 2006-07-04 Triogo, Inc. Process for improving search engine efficiency using feedback
AU5934900A (en) 1999-07-16 2001-02-05 Agentarts, Inc. Methods and system for generating automated alternative content recommendations
US6421672B1 (en) * 1999-07-27 2002-07-16 Verizon Services Corp. Apparatus for and method of disambiguation of directory listing searches utilizing multiple selectable secondary search keys
US6567786B1 (en) 1999-09-16 2003-05-20 International Business Machines Corporation System and method for increasing the effectiveness of customer contact strategies
US6839680B1 (en) 1999-09-30 2005-01-04 Fujitsu Limited Internet profiling
US6681247B1 (en) 1999-10-18 2004-01-20 Hrl Laboratories, Llc Collaborator discovery method and system
US6489968B1 (en) 1999-11-18 2002-12-03 Amazon.Com, Inc. System and method for exposing popular categories of browse tree
GB2358263A (en) 2000-01-13 2001-07-18 Applied Psychology Res Ltd Generating user profile data
US6721744B1 (en) 2000-01-28 2004-04-13 Interval Research Corporation Normalizing a measure of the level of current interest of an item accessible via a network
WO2001069452A2 (en) 2000-03-14 2001-09-20 Blue Dolphin Group, Inc. Method of selecting content for a user
US6389469B1 (en) 2000-03-27 2002-05-14 Targetize Innovative Solutions Ltd. System and method for customized content delivery
US6484164B1 (en) 2000-03-29 2002-11-19 Koninklijke Philips Electronics N.V. Data search user interface with ergonomic mechanism for user profile definition and manipulation
AU2001277071A1 (en) 2000-07-21 2002-02-13 Triplehop Technologies, Inc. System and method for obtaining user preferences and providing user recommendations for unseen physical and information goods and services
US6882977B1 (en) 2000-07-31 2005-04-19 Hewlett-Packard Development Company, L.P. Method and facility for displaying customer activity and value
US20020116258A1 (en) 2000-12-06 2002-08-22 George Stamatelatos Method for selecting and directing internet communications
US20030074400A1 (en) 2001-03-30 2003-04-17 David Brooks Web user profiling system and method
US20030023481A1 (en) 2001-07-24 2003-01-30 Sarah Calvert Method of selecting an internet advertisement to be served to a user
US20030078928A1 (en) 2001-10-23 2003-04-24 Dorosario Alden Network wide ad targeting
US7130866B2 (en) * 2002-07-30 2006-10-31 Koninklijke Philips Electronics N.V. Controlling the growth of a feature frequency profile by deleting selected frequency counts of features of events
US20040024719A1 (en) 2002-07-31 2004-02-05 Eytan Adar System and method for scoring messages within a system for harvesting community kowledge

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6415322B1 (en) * 1998-02-27 2002-07-02 Engage, Inc. Dual/blind identification
US6421675B1 (en) * 1998-03-16 2002-07-16 S. L. I. Systems, Inc. Search engine
US20020111994A1 (en) * 2001-02-14 2002-08-15 International Business Machines Corporation Information provision over a network based on a user's profile
US20030128236A1 (en) * 2002-01-10 2003-07-10 Chen Meng Chang Method and system for a self-adaptive personal view agent
US20070061195A1 (en) * 2005-09-13 2007-03-15 Yahoo! Inc. Framework for selecting and delivering advertisements over a network based on combined short-term and long-term user behavioral interests

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
'Proceedings of the 13th international conference on World Wide Web, May 17-20, 2004, New York, NY, USA', article KAZUNARI SUGIYAMA ET AL.: 'Adaptive web search based on user profile constructed without any effort from users' *
'Proceedings of the eighth international conference on Information and knowledge management, p.405-412, November 02-06, 1999, Kansas City, Missouri, United States', article DWI H. WIDYANTORO: 'An adaptive algorithm for learning changes in user interests', pages 405 - 412 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012024066A2 (en) * 2010-08-19 2012-02-23 Yahoo! Inc. Contextual advertising with user features
WO2012024066A3 (en) * 2010-08-19 2012-05-24 Yahoo! Inc. Contextual advertising with user features
US9031863B2 (en) 2010-08-19 2015-05-12 Yahoo! Inc. Contextual advertising with user features

Also Published As

Publication number Publication date
US7734632B2 (en) 2010-06-08
US20100250558A1 (en) 2010-09-30
US8131733B2 (en) 2012-03-06
WO2007055812A3 (en) 2009-04-23
US20070100796A1 (en) 2007-05-03

Similar Documents

Publication Publication Date Title
US7734632B2 (en) System and method for targeted ad delivery
US20190279230A1 (en) Online Content Delivery Based on Information from Social Networks
JP6441817B2 (en) Start real-time bidding based on expected bidding revenue
JP5053483B2 (en) Network for retargeted ad distribution
US7647387B2 (en) Methods and systems for rule-based distributed and personlized content delivery
US7739708B2 (en) System and method for revenue based advertisement placement
US8484073B2 (en) Method of distributing targeted internet advertisements
US20150235258A1 (en) Cross-device reporting and analytics
US7870227B2 (en) System and method for merging internet protocol address to location data from multiple sources
US11386450B2 (en) Systems and methods for advertising on content-screened web pages
US9767488B1 (en) Bidding based on the relative value of identifiers
US10621600B2 (en) Method for analyzing website visitors using anonymized behavioral prediction models
US20120290373A1 (en) Apparatus and method for marketing-based dynamic attribution
US20100223105A1 (en) Method and apparatus for content targeting one user group based on behavioral profiling of another user group
JP2008524701A (en) Audience harmony network for performance disaggregation and revenue allocation
JP2008524700A (en) Audience harmony network for performance disaggregation and revenue allocation
KR20110137289A (en) Determining conversion probability using session metrics
WO2010132571A2 (en) Methods and apparatus for portfolio and demand bucket management across multiple advertising exchanges
US20100257022A1 (en) Finding Similar Campaigns for Internet Advertisement Targeting
KR20090089919A (en) Network node ad targeting
JP2016517094A (en) Systems and methods for audience targeting
EP2891995A1 (en) Systems and methods for search results targeting
US20080082655A1 (en) Click Value Determination With Incentive Schemes For Website Visitors And Advertisers
US20150066644A1 (en) Automated targeting of information to an application user based on retargeting and utilizing email marketing
US20150287096A1 (en) Bid control through semantic signals expressed in natural language

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 06815467

Country of ref document: EP

Kind code of ref document: A2