Computer processes for adaptively selecting and/or ranking items for display in particular contexts

• US 7,779,014 B2
• Filed: 10/19/2007
• Issued: 08/17/2010
• Est. Priority Date: 10/30/2001
• Status: Active Grant

• Current Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Original Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Inventors: York, Jeremy C., Smith, Brent, Cottrell, Robert
• Primary Examiner(s): Ali; Mohammad
• Assistant Examiner(s): Shanmugasundaram; Kannan
• Application Number:
  US11/875,757
• Publication Number:
  US 20080040341A1
• Time in Patent Office: 1,033 Days
• Field of Search: None
• US Class Current: 707/748
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0633 Lists, e.g. purchase orders, compilation or processing
• CPC Class Codes: G06Q 30/0641 Shopping interfaces
• CPC Class Codes: Y10S 707/948 Product or catalog
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A computer-implemented system and method are disclosed that are capable of refining the order and/or content of a list of items. In one embodiment, user activity associated with each item displayed to users in a given context is monitored, and is used together with item exposure data to generate context-specific item weights. The weights for the same item in different contexts may be different. The selection of items included in a list presented to users, and/or the order in which the items in the list are displayed, may be adjusted over time based upon the relative weights associated with the items.

Claims

1. A computer-implemented method, comprising:

tracking item exposure events in which items are exposed to users on web pages in a plurality of different contexts, wherein tracking the item exposure events comprises (a) identifying the context of each item exposure event based on one or more user actions leading to the exposure event, and (b) recording each item exposure event in computer storage in association with the identified context;

recording user activity resulting from the item exposure events, said user activity comprising item selection events in which a user selects an item exposed on a web page, wherein recording the user activity comprises recording each item selection event in computer storage in association with the context of the corresponding item exposure event;

receiving a request from a browser with a server, said request corresponding to a particular context of the plurality of contexts;

identifying a plurality of items that are responsive to the request, at least some of said items capable of being displayed in more than one of said contexts;

calculating context-specific weights for each of the items, each contextspecific weight corresponding to a respective item of said plurality of items and being specific to said particular context, each context-specific weight being dependent upon a number of times the respective item has been exposed in said particular context, as determined from said tracking of item exposure events, and additionally being dependent upon the recorded user activity that has resulted from said exposure of the respective item in said context, wherein the contextspecific weights are dependent upon item selection actions of a plurality of users;

generating a ranked list of said plurality of items such that positions of said items in the ranked list are dependent upon said context-specific item weights;

generating and outputting a web page in response to the request, said web page including at least a portion of the ranked list;

recording at least one item selection action performed by a user on said web page; and

adjusting at least one of said weights by an amount that is dependent upon the at least one item selection action.

2. The method of claim 1, wherein the method comprises selecting a subset of the items in the ranked list to display on the web page, such that less than all of said plurality of items are included on said web page.

3. The method of claim 2, further comprising:

recording exposure data for said request, said exposure data identifying the subset of items selected to include on the web page;

wherein said amount by which at least one of the weights is adjusted is additionally dependent upon the exposure data.

4. The method of claim 1, wherein identifying said plurality of items comprises identifying items that are responsive to criteria specified in the request.

5. The method of claim 1, wherein the weights are non-user-specific.

6. The method of claim 1, wherein the web page comprises user-selectable display elements for initiating a display of additional items that are similar to, but different from, the items on the web page, and the method further comprises monitoring users selections of said display elements and resulting user selections of the additional items, and based thereon, causing an additional item to be added to a list of items associated with the particular context.

7. A computer-implemented method of adaptively controlling the display of an item, the method comprising:

for each of a plurality of contexts in which the item is capable of being displayed, tracking both context-specific exposure of the item and contextspecific user activity resulting from said exposure, wherein said tracking comprises storing context-specific exposure data and context-specific user activity data in computer storage for each of the plurality of contexts, said context-specific exposure data reflecting item exposure events in which the item is included on dynamically generated pages served to users by a server system, said context-specific user activity data reflecting item selection events in which users select the item as a result of its exposure on said dynamically generated pages;

for each of the plurality of contexts, generating a respective contextspecific item weight for the item, said context-specific item weight being dependent upon the context-specific exposure data and the context-specific user activity data corresponding to the respective context, and not being specific to any particular user, wherein generating the context-specific item weights comprises adjusting a current value of a context-specific item weight to reflect the context-specific exposure data and context-specific user activity data resulting from said value;

using the context-specific item weights to at least rank the item, relative to other items, for display to users in the respective contexts to which such item weights correspond;

causing the item to be displayed, in a first context of said plurality of contexts, together with a user-selectable option for initiating a display of additional items that are similar to said item;

monitoring user selections of said option, and resulting user selections of the additional items; and

when said user selections of additional items reveal a threshold degree of association between an additional item and said first context, causing the additional item to be added to a list of items associated with the first context.

8. The method of claim 7, wherein the context-specific exposure data reflects a number of times the item has been included in pages generated by said server system, and the context-specific user activity data reflects a number of times the item has been selected by users from said pages.

9. The method of claim 7, wherein generating the context-specific item weights comprises applying a time-based decay function to the exposure data and the user activity data to cause greater weight to be given to recent exposure and user activity data than to older exposure and user activity data.

10. The method of claim 7, wherein tracking the contextspecific exposure of the item comprises separately maintaining, for each of said contexts, a count of a number of times the item has been selected for inclusion in pages served to users.

11. The method of claim 7, wherein each context-specific item weight is dependent upon tracked activity of a plurality of users.

12. The method of claim 7, wherein each of the contexts corresponds to a respective list of items.

13. The method of claim 7, wherein each of the contexts corresponds to a respective task capable of being performed by users.

14. The method of claim 7, further comprising:

receiving a page request from a browser, said page request corresponding to said first context of said plurality of contexts; and

in response to the page request, generating a page in which a display position of the item, relative to other items on the page, is dependent upon the context-specific item weight corresponding to the first context.

15. The method of claim 7, wherein the item is a catalog item in an electronic catalog.

16. The method of claim 7, wherein the method is performed to control a display of the item on web pages of a web site, and the item is a link to another web site.

17. The method of claim 7, wherein tracking the contextspecific exposure of the item comprises identifying the context associated with an item exposure event based on a sequence of user actions which led to the item exposure event.

18. The method of claim 17, wherein the sequence of user actions comprises user responses to a series of questions.

19. A system for adaptively selecting items to present to users, comprising:

a server system that is responsive to page requests received over a network by identifying contexts associated with said page requests, and by dynamically generating and outputting pages that include ordered displays of items, wherein the server system identifies a context associated with a page request based at least partly on one or more user actions leading to said page request;

an exposure tracking system that tracks, for each of a plurality of items, item exposure events resulting from inclusion of the item on the dynamically generated pages, including item exposure events that do not result in user selection of the item, wherein the exposure tracking system tracks the item exposure events separately for each of a plurality of said contexts by recording the item exposure events in computer storage in association with the identified contexts;

an activity tracking system that tracks item selection actions performed by users in response to the item exposure events, wherein the activity tracking system tracks the item selection events separately for each of the plurality of contexts; and

an item weighting system that generates, for at least a first item of the plurality of items, a plurality of context-specific item weights, each of which corresponds to a respective context, said context-specific item weights being dependent upon the item exposure events and item selection events tracked by the exposure tracking system and the activity tracking system, respectively, said context-specific item weights being dependent upon item selection actions of a plurality of users, wherein the item weighting system adjusts current values of the context-specific item weights by amounts that depend upon the tracked item selection actions resulting from said current values;

wherein the server system uses the context-specific item weights for the first item to rank the first item relative to the other items, and to select the first item for display, in the respective contexts to which such item weights correspond.

20. The system of claim 19, wherein each context-specific item weight is dependent upon tracked activity of a plurality of users.

21. The system of claim 19, wherein the item exposure events tracked by the exposure tracking system are a result of the server system using the context-specific item weights to select items for display.

22. The system of claim 19, wherein the server system uses the context-specific item weights to control a display of the first item in responses to search requests.

23. The system of claim 19, wherein the server system uses each context-specific item weight to control a display of the first item to a plurality of users.

24. The system of claim 19, wherein the items include identifiers of products represented in an electronic catalog.

25. The system of claim 19, wherein the items include links to web sites.

26. The system of claim 19, wherein the exposure tracking system is operative to identify the context associated with an item exposure event based on a sequence of user actions which led to the item exposure event.

27. The method of claim 26, wherein the sequence of user actions comprises user responses to a series of questions.