Providing ad information using plural content providers

• US 7,882,046 B1
• Filed: 03/30/2007
• Issued: 02/01/2011
• Est. Priority Date: 11/10/2006
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Haberkorn, Karen R., Perkowitz, Mike, Dickinson, Andrew B., Mason, Llewllyn J., Cole, Richard J., Schuster, Brian J., Kizhakkekalathil, Raj K., Jacobson, Mark
• Primary Examiner(s): Holmes; Michael B.
• Assistant Examiner(s): GONZALES, VINCENT
• Application Number:
  US11/694,699
• Time in Patent Office: 1,404 Days
• Field of Search: 705/14, 706/12
• US Class Current: 706/12
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06Q 30/0254 based on statistics

Abstract

An ad system is described for providing ad selections in response to an ad presentation opportunity. The ad system can use multiple content providers to generate multiple sets of ad candidates. The content providers can apply different ad selection paradigms in generating their sets of ad candidates. The paradigms may act on different aspects of a context pertaining to the ad presentation opportunity. A chooser module and filtering module can cull the set of ad selections from among the plural sets of ad candidates. At least one content provider uses machine learning functionality in generating ad candidates, such as a Naïve Bayesian approach, an affiliate similarity approach, etc. Various content providers also find application in a stand-alone mode.

Claims

1. An ad-serving system executable on one or more processors for providing ad selections, comprising:

an item management module for offering items to users, the item management module including a data store for maintaining a catalog of the items being offered;

a plurality of content providers each having a plurality of advertisements and each advertisement being associated with a unique identifier,

a plurality of affiliate modules for interacting with the users, each affiliate module having one or more tag identifiers; and

an ad system for generating ad selections which are associated with the items being offered by the item management module, the ad selections being presentable to the users in the context of electronic services provided by the plurality of affiliate modules,

wherein the ad system comprises;

an ad serving module for providing ad selections to the users; and

an ad selection module for generating ad selections to be delivered to the users, wherein the ad selection module is operative to generate the ad selections by choosing from plural sets of ad candidates, the choosing based at least in part on a ranking of the unique identifiers for a particular one of the tag identifiers, each set being obtained from a respective one of the plurality of content providers and the plural set of ad candidates being provided using plural respective ad selection paradigms, wherein the ad selection module is operative to generate the ad selections based on at least one aspect of an ad presentation context, said at least one aspect comprising text adjacent to an ad space designated for presentation of an ad.

2. The system of claim 1, wherein said least one aspect of the ad presentation context further comprises one or more of:

an aspect that pertains to one or more characteristics of content with which the users interact;

an aspect that pertains to one or more characteristics of the affiliate modules;

or

an aspect that pertains to one or more characteristics of the users.

3. A computerized method for generating one or more ad candidates, comprising:

in response to an ad presentation opportunity at an affiliate component, receiving plural sets of ad candidates using plural respective ad selection paradigms, each set being received from a respective one of a plurality of content providers and each set comprising a tag identifier associated with the affiliate component; and

choosing one or more of the ad candidates each having a unique identifier from the plural sets of the ad candidates to provide one or more chosen ad candidates, the choosing based at least in part on text adjacent to one or more ad spaces designated for presentation of the chosen ad candidates.

4. The computerized method of claim 3, further comprising forwarding said one or more chosen ad candidates to a user.

5. The computerized method of claim 3, further comprising filtering said one or more chosen ad candidates based on at least one filtering criterion to provide one or more ad selections.

6. The computerized method of claim 5, wherein said at least one filtering criterion directs the removal of one or more chosen ad candidates associated with items that have an undesirable availability status.

7. The computerized method of claim 5, wherein said at least one filtering criterion directs the removal of one or more chosen ad candidates that are assessed as duplicates.

8. The computerized method of claim 5, wherein said at least one filtering criterion directs the removal of one or more chosen ad candidates having undesirable content.

9. The computerized method of claim 3, wherein the choosing further comprises selecting at least one creative part that governs a manner in which ad information is presented to a user.

10. The computerized method of claim 9, wherein at least one creative part provides a directive to present a single featured item part.

11. The computerized method of claim 9, wherein at least one creative part provides a directive to present plural featured item parts.

12. The computerized method of claim 3, wherein the generating of plural sets of ad candidates is based on one or more aspects of a context associated with the ad presentation opportunity.

13. The computerized method of claim 12, wherein an aspect of the context pertains to one or more characteristics of content with which a user is interacting.

14. The computerized method of claim 12, wherein an aspect of the context pertains to one or more characteristics of an affiliate module with which a user is interacting.

15. The computerized method of claim 12, wherein an aspect of the context pertains to one or more characteristics of a user.

16. The computerized method of claim 3, further comprising recording a user'"'"'s action in response to a presentation of said one or more chosen ad candidates.

17. The computerized method of claim 3, wherein a model is used to generate at least one of the plural sets of ad candidates, further comprising modifying the model in response to feedback concerning actions taken by users in response to being served ad candidates.

18. The computerized method of claim 3, wherein a model is used to perform the choosing, further comprising modifying the model in response to feedback concerning actions taken by users in response to being served ad candidates.

19. The computerized method of claim 18, wherein the choosing uses a Bayesian model to generate said one or more chosen ad candidates.

20. The computerized method of claim 19, wherein the Bayesian model uses a Naï

ve Bayesian classification paradigm.

21. The computerized method of claim 3, wherein the chosen ad candidates are presented to a user in the form of at least one ad unit, further comprising receiving a configuration selection that customizes a characteristic of said at least one ad unit.

22. The computerized method of claim 21, wherein the received configuration selection governs an appearance of said at least one ad unit.

23. The computerized method of claim 21, wherein the received configuration selection governs information imparted by said at least one ad unit.

24. The computerized method of claim 21, wherein the received configuration selection governs a functional behavior of said at least one ad unit.

25. An ad system for generating ad candidates, comprising:

logic executed on one or more processors to identify an ad presentation opportunity, the ad presentation opportunity having a context, wherein the context has one or more aspects selected from one or more characteristics of content with which a user is interacting, wherein at least one of the one or more characteristics comprises text adjacent to an ad space in the content designated by the ad presentation opportunity; and

logic executed on one or more processors to generate ad candidates from sets of ad candidates received from a plurality of content providers using plural ad selection paradigms, wherein different ad selection paradigms provide ad candidates that are associated with different respective aspects of the context.

26. The ad system of claim 25, wherein the one or more characteristics of the content further comprises one or more characteristics of an affiliate module with which the user is interacting.

27. The ad system of claim 25, wherein the one or more characteristics of the content further comprises one or more characteristics of the user.

28. A computer-implemented method for choosing one or more ad candidates from plural sets of ad candidates, comprising:

under control of one or more computer systems configured with executable instructions,

receiving a request for an ad selection from an affiliate component, the request comprising a tag identifier associated with the affiliate component;

receiving plural sets of ad candidates from a plurality of content providers and generated based on respective different ad selection paradigms, each ad candidate having a unique identifier and each of the plural sets of ad candidates being received from a respective one of a plurality of content providers;

deriving weights from a unit vector matrix comprising the tag identifiers and the unique identifiers;

applying the weights to the respective ad selection paradigms; and

choosing one or more ad candidates from the plural sets of ad candidates based at least in part on the weights applied to the ad selection paradigms, text adjacent to an ad space designated for presentation of an ad, and a sales commitment for each of the plurality of affiliate components.

29. The computerized method of claim 28, further comprising generating the weights using a model, wherein the model is derived based at least in part on past performance of ad selections served to users for different ad presentation opportunities.

30. The computerized method of claim 29, wherein the model uses a Bayesian approach to generate the weights.

31. One or more computer-readable storage media containing computer-readable instructions, the instructions implementing an ad generation component when executed by one or more processors, the computer readable instructions comprising:

logic operative to, in response to an ad presentation opportunity from an affiliate component, receive from a plurality of content providers plural sets of ad candidates from a plurality of content providers, wherein each ad candidate has a unique identifier and each of the plurality of content providers uses plural respective ad selection paradigms;

logic operative to choose one or more of the ad candidates from the plural sets of the ad candidates to provide one or more chosen ad candidates, the logic being further operative to choose based at least in part upon text adjacent to an ad space designated for presentation of an ad; and

logic operative to filter said one or more chosen ad candidates based on at least filtering criteria comprising a dynamic revenue share configured to facilitate achievement of a pre-determined earnings level of one or more of the plurality of affiliate components.

32. The one or more computer-readable storage media of claim 31, wherein the logic operative to generate plural sets of ad candidates is operative to generate the sets based on one or more aspects of a context associated with the ad presentation opportunity.

33. The one or more computer-readable storage media of claim 32, wherein an aspect of the context pertains to one or more characteristics of content with which a user is interacting.

34. The one or more computer-readable storage media of claim 32, wherein an aspect of the context pertains to one or more characteristics of an affiliate module with which a user is interacting.

35. The one or more computer-readable storage media of claim 32, wherein an aspect of the context pertains to one or more characteristics of a user.