Computer-based analysis of affiliate web site performance

• US 20060106665A1
• Filed: 11/12/2004
• Published: 05/18/2006
• Est. Priority Date: 11/12/2004
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Kumar, Dilip S., Rusmevichientong, Paat, Anderson, Michelle K., Selinger, David L.
• Granted Patent: US 7,797,197 B2
• Time in Patent Office: Days
• US Class Current: 705/10
• CPC Class Codes: G06Q 10/06393 Score-carding, benchmarking or key performance indicator [KPI] analysis
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0211 Determining the effectiveness of discounts or incentives
• CPC Class Codes: G06Q 30/0241 Advertisements
• CPC Class Codes: G06Q 30/0244 Optimization
• CPC Class Codes: G06Q 30/0247 Calculate past, present or future revenues
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G06Q 30/0631 Item recommendations

Abstract

A performance analysis engine analyzes the performance of affiliate web sites that provide links to specific items in an electronic catalog, and identifies catalog items that can be listed by such affiliate web sites to improve performance. An association mining component analyzes transaction data attributable to specific categories or types of affiliate web sites to identify items that are frequently purchased in combination by users of such sites. The detected item associations are used to evaluate, for a given affiliate site, whether significant disparities exist between the expected and actual sales quantities of specific items. The results of the analysis are incorporated into affiliate-specific performance reports, which may include specific recommendations for improving performance. The disclosed methods may also be used to analyze the performance of, and provide recommendations to, online sellers within an online mall, online marketplace, or online auction system.

Claims

1. A computer-implemented method of analyzing the performance of affiliate sites that refer users to an online commerce site, the method comprising:

storing transaction data descriptive of purchases made by users referred to the online commerce site by a plurality of affiliate sites;

programmatically applying an association-rule mining algorithm to the transaction data to generate a set of association rules, said association rules identifying items that tend to be purchased in combination by said users; and

programmatically analyzing performance of a particular affiliate site by using the association rules, in combination with transaction data associated with the particular affiliate site, to evaluate whether disparities exist between expected and actual quantities of specific items purchased by users referred by the particular affiliate site.

2. The method of claim 1, wherein each association rule identifies a first set of items, X1, and a second set of items, X2, and specifies a conditional probability that a purchaser of X1 will also purchase X2.

3. The method of claim 1, wherein each association rule comprises a data value representing a conditional probability that a purchaser of X1 will also purchase X2, where X1 and X2 are mutually exclusive sets of one or more items.

4. The method of claim 1, wherein at least one of the association rules maps a two-item set to a single item.

5. The method of claim 1, wherein the set of association rules is specific to a particular affiliate site category in which the particular affiliate site falls, such that the performance of the particular affiliate site is compared to the performance of other affiliate sites within said category.

6. The method of claim 5, wherein programmatically applying an association-rule mining algorithm comprises using category information supplied by affiliates to identify affiliate sites falling within said category.

7. The method of claim 1, wherein the association rules are specific to a particular type of affiliate site, and are generated based on transaction data attributable to affiliate sites of said type.

8. The method of claim 1, further comprising generating an electronic report reflective of one or more disparities detected by programmatically analyzing the performance of the particular affiliate site, and transmitting the report to a corresponding affiliate.

9. The method of claim 8, wherein the electronic report includes at least one recommendation of an item to be listed on the particular affiliate site.

10. The method of claim 8, wherein the electronic report includes link coding for adding an item-specific affiliate link to the particular site.

11. A computer system programmed to perform the method of claim 1.

12. A computer-readable medium that stores a computer program which, when executed by a computer system, causes the computer system to perform the method of claim 1.

13. A system for analyzing the performance of affiliate sites, the system comprising:

a data repository that stores transaction data descriptive of item purchases made from an electronic catalog by users referred by the affiliate sites;

an association mining component that analyzes, and generates association rules reflective of, the transaction data stored in the data repository, said association rules identifying items that are associated based on purchase activities of referred users; and

a disparity detection module that uses the association rules generated by the association mining component, in combination with transaction data descriptive of item purchases made by users referred by a particular affiliate site, to evaluate whether disparities exist between expected and actual performance of the particular affiliate site.

14. The system of claim 13, wherein the disparity detection module evaluates whether disparities exist between expected and actual quantities of specific items purchased by users referred by the particular affiliate site.

15. The system of claim 13, wherein the association mining component generates an association rule that (a) identifies a first set, X1, of one or more items, and a second set, X2, of one or more items, and (b) specifies a conditional probability that a purchaser of X1 will also purchase X2, and wherein the disparity detection module uses the association rule to evaluate performance of the particular affiliate site with respect to an item in X2.

16. The system of claim 13, wherein the association mining component generates association rules comprising data values representing conditional probabilities that purchasers of X1 will also purchase X2, where X1 and X2 are sets of one or more items.

17. The system of claim 13, wherein the association mining component generates an association rule that maps a two-item set to a single item, and the disparity detection module uses the association rule to evaluate performance of the particular affiliate site with respect to the single item.

18. The system of claim 13, wherein the association mining component generates a set of association rules that is specific to an affiliate site category in which the particular affiliate site falls, and the disparity detection module uses the set of association rules to evaluate performance of the particular affiliate site relative to performance of other affiliate sites in said category.

19. The system of claim 18, wherein the association mining component uses category information supplied by affiliates to identify affiliate sites falling within said category.

20. The system of claim 13, further comprising a report generation module that generates reports reflective of disparities detected by the disparity detection module, and which transmits the reports to corresponding affiliates.

21. The system of claim 20, wherein the reports include recommendations of items to list.

22. A computer-implemented method of analyzing the performance of affiliate sites that refer users to an online commerce site, the method comprising:

programmatically generating, for each of a plurality of categories of affiliate sites, a respective set of item association data that identifies items that are purchased in combination relatively frequently by users referred by affiliate sites falling within the respective category; and

programmatically analyzing performance of a particular affiliate site that falls within a first category of said plurality of categories by comparing the set of item association data for said first category to transaction data associated with the particular affiliate site.

23. The method of claim 22, wherein programmatically analyzing the performance of the particular affiliate site comprises identifying items to add to the particular affiliate site to improve said performance.

24. The method of claim 22, wherein programmatically analyzing the performance of the particular affiliate site comprises evaluating whether disparities exist between expected and actual quantities of specific items purchased by users referred by the particular affiliate site.

25. The method of claim 24, further comprising generating an electronic report reflective of one or more disparities detected by programmatic analysis of the performance of the particular affiliate site, and transmitting the report to a corresponding affiliate.

26. The method of claim 25, wherein the report includes at least one recommendation of an item to be listed on the particular affiliate site.

27. The method of claim 22, wherein each set of item association data comprises one or more association rules.

28. The method of claim 27, wherein each association rule identifies a first set of on items, X1, and a second set of items, X2, and specifies a conditional probability that a purchaser of X1 will also purchase X2.

29. The method of claim 27, wherein at least one of the association rules maps a two-item set to a single item.

30. The method of claim 22, wherein the item association data comprises data values reflective of frequencies with which specific items are purchased in combination.

31. The method of claim 22, wherein generating the item association data comprises using category information supplied by affiliates to identify affiliate sites falling within each of the plurality of categories.

32. The method of claim 22, wherein generating the item association data for each category comprises programmatically categorizing the affiliate sites based on at least one of (a) content of the affiliate sites, and (b) purchases made by users referred by the affiliate sites.

33. A computer system programmed to perform the method of claim 22.

34. A computer-readable medium that stores a computer program which, when executed by a computer system, causes the computer system to perform the method of claim 22.

35. A computer-implemented method of analyzing the performance of online sellers, comprising:

programmatically generating, for each of a plurality of categories of online sellers, a respective set of item association rules that reflect item purchase patterns of users that have purchased items from one or more online sellers in the respective category; and

programmatically analyzing performance of a particular online seller that falls within a first category of said plurality of categories by comparing the set of item association rules for said first category to transaction data associated with the particular online seller.

36. The method of claim 35, wherein the set of item association rules for a given category identifies items that are purchased in combination relatively frequently by customers of the online sellers falling within the given category.

37. The method of claim 35, wherein programmatically generating a respective set of item association rules comprises analyzing aggregated item sales data of a plurality of online sellers falling within the respective category.

38. The method of claim 35, wherein programmatically analyzing the performance of the particular online seller comprises identifying items to recommend to the particular online seller.

39. The method of claim 35, wherein programmatically analyzing the performance of the particular online seller comprises evaluating whether disparities exist between expected and actual quantities of specific items sold by the particular online seller.

40. The method of claim 39, further comprising generating an electronic report reflective of one or more disparities detected by programmatic analysis of the performance of the particular online seller, and transmitting the report to the particular online seller.

41. The method of claim 40, wherein the report includes at least one recommendation of an item to be listed by the particular online seller.

42. The method of claim 35, wherein each item association rule identifies a first set of on items, X1, and a second set of items, X2, and specifies a conditional probability that a purchaser of X1 will also purchase X2.

43. The method of claim 35, wherein at least one of the item association rules maps a two-item set to a single item.

44. The method of claim 35, wherein generating the item association rules comprises using category information supplied by the online sellers to identify online sellers falling within each of the plurality of categories.

45. The method of claim 35, wherein the online sellers include merchants in an electronic mall system.

46. The method of claim 35, wherein the online sellers include sellers in at least one of (a) an electronic marketplace, and (b) an online auction system.

47. The method of claim 35, wherein the plurality of online sellers include affiliates that list specific items for sale on affiliate web sites.

48. A computer system programmed to perform the method of claim 35.

49. A computer-readable medium that stores a computer program which, when executed by a computer system, causes the computer system to perform the method of claim 35.