Use of product viewing histories of users to identify related products

• US 20020019763A1
• Filed: 03/29/2001
• Published: 02/14/2002
• Est. Priority Date: 09/18/1998
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon.com, Inc.
• Inventors: Smith, Brent R., Linden, Gregory D.
• Granted Patent: US 6,912,505 B2
• Time in Patent Office: Days
• US Class Current: 705/10
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0253 During e-commerce, i.e. online transactions
• CPC Class Codes: G06Q 30/0254 based on statistics
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0271 Personalized advertisement
• CPC Class Codes: G06Q 30/0277 Online advertisement
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G06Q 30/0603 Catalogue ordering
• CPC Class Codes: G06Q 30/0631 Item recommendations
• 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: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

Various methods are disclosed for monitoring user browsing activities that indicate user interests in particular products or other items, and for using such information to identify items that are related to one another. In one embodiment, relationships between products within an online catalog are determined by identifying products that are frequently viewed by users within the same browsing session (e.g., products A and B are related because a significant portion of those who viewed A also viewed B). The resulting item relatedness data is preferably stored in a table that maps items to sets of related items. The table may be used to provide personalized product recommendations to users, and/or to supplement product detail pages with lists of related products. In one embodiment, the table is used to provide session-specific product recommendations to users that are based on the products viewed by the user during the current browsing session.

Claims

1. A computer-implemented method of using online browsing activities of users to identify related products, comprising:

generating, for each of a plurality of users, a history of products viewed by the user during online browsing, to thereby generate a plurality of user-specific product viewing histories;

analyzing the plurality of product viewing histories to identify correlations between product viewing events reflected therein; and

generating a data structure that maps each product of a plurality of products to a respective set of related products, wherein product relationships indicated by the data structure reflect said correlations between product viewing events.

2. The method of claim 1, wherein each of the plurality of product viewing histories is specific to a single browsing session.

3. The method of claim 1, wherein analyzing the plurality of product viewing histories to identify correlations comprises determining frequencies with which specific products co-occur within the product viewing histories of the users.

4. The method of claim 1, wherein analyzing the plurality of product viewing histories comprises, for each of a plurality of product pairs of a first product and a second product, determining a number of product viewing histories in which both the first product and the second product occur.

5. The method of claim 1, wherein analyzing the plurality of product viewing histories comprises, for each of a plurality of product pairs of a first product and a second product, determining a number of browsing sessions in which both the first product and the second product were viewed.

6. The method of claim 1, wherein at least some of the product viewing histories respectively reflect product viewing events over multiple sessions by the same user.

7. The method of claim 1, wherein the product viewing events include user accesses to product information that is output to users solely by voice.

8. The method of claim 1, wherein generating a history of products viewed comprises recording product viewing events within a session-specific record during the browsing session.

9. The method of claim 8, further comprising using the session-specific record to provide personal recommendations to the user during the session.

10. The method of claim 1, wherein generating a history of products viewed comprises treating a user access to a detail page of a product as viewing of that product.

11. The method of claim 1, wherein generating a history of products viewed comprises identifying products displayed on at least one of a search results page and a browse node page.

12. The method of claim 1, further comprising using the data structure to provide personalized recommendations to a visitor by at least:

identifying multiple products that are of known interest to the visitor;

for each of the products of known interest, obtaining from the data structure a corresponding set of related products;

combining the sets of related products to generate a combined set of related products; and

selecting at least a portion of the combined set of related products to recommend to the visitor.

13. The method of claim 12, wherein identifying products that are of known interest to the visitor comprises identifying a set of products viewed by the visitor during a current browsing session, to thereby generate recommendations that are specific to the browsing session.

14. The method of claim 12, wherein identifying products that are of known interest to the visitor comprises treating a visit by the user to a detail page of a product as an indication that the visitor is interested in the product.

15. The method of claim 1, further comprising:

(a) accessing the data structure to identify a set of products that are related to a first product; and

(b) displaying at least some of the products that are related to the first product on a product detail page of the first product.

16. The method as in claim 15, wherein (a) is performed in real time in response to a user request for the product detail page.

18. The method of claim 17, wherein each user-specific product viewing history is specific to a respective browsing session.

19. The method of claim 17, wherein (b) comprises determining a frequency with which both the first and second products were viewed by users within the same browsing session.

20. The method of claim 17, wherein at least some of the user-specific product viewing histories reflect browsing activities of users over multiple web sites.

21. The method of claim 17, further comprising:

(d) comparing degrees of relatedness determined in (b) and (c) to identify a subset of the second products that are deemed most closely related to the first product.

22. The method of claim 21, further comprising performing (b)-(d) for each of a plurality of additional first products, and storing resulting data in a data structure that maps each first product to a respective set of related products.

23. The method of claim 22, further comprising using the data structure to provide personalized recommendations to a user by at least:

identifying multiple products that are of known interest to the user;

for each of the products of known interest, identifying from the data structure a corresponding set of related products;

combining the sets of related products to generate a combined set of related products; and

selecting at least a portion of the combined set of related products to recommend to the user.

24. The method of claim 23, wherein identifying products that are of known interest to the user comprises identifying a set of products viewed by the user during a current browsing session such that the recommendations are specific to the browsing session.

25. The method of claim 22, further comprising:

(a) accessing the data structure to identify a set of products that are related to a first product; and

(b) displaying at least some of the products that are related to the first product on a product detail page of the first product.

26. The method as in claim 25, wherein (a) is performed in real time in response to a user request for the product detail page.

28. The method as in claim 27, wherein (a) comprises treating a user access to a detail page of an item as viewing of that item.

29. The method as in claim 27, wherein the items are products within an online catalog.

30. The method as in claim 29, wherein (a) comprises treating a user access to a detail page of a product as viewing of that product.

31. The method as in claim 27, further comprising comparing degrees of relatedness determined in (b) and (c) to identify a set of second items that are most closely related to the first item.

17. A computer-implemented method of identifying products that are related to a first product, comprising:

(a) generating a plurality of user-specific product viewing histories for a plurality of users, wherein each product viewing history indicates products viewed by a user during online browsing of a collection of products;

(b) determining a degree of relatedness between the first product and a second product, such that the degree of relatedness reflects a frequency with which both the first and second products occur within the same product viewing history of the plurality of user-specific product viewing histories; and

(c) performing (b) for each of a plurality of additional second products.

27. A computer-implemented method of identifying a set of items that are related to a first item, comprising:

(a) generating, for each of a plurality of browsing sessions, a history of items viewed during the browsing session, to thereby generate a plurality of session-specific item viewing histories;

(b) determining a degree of relatedness between the first item and a second item such that the degree of relatedness reflects a frequency with which the first and second items were viewed within the same browsing session as reflected within the plurality of session-specific item viewing histories; and

(c) performing (b) for each of a plurality of additional second items.

32. A method of supplementing product detail pages within an online catalog of products, comprising:

processing product viewing histories of a plurality of users to identify, for a first product, a set of additional products that are related to the first product based at in part on co-occurrences of each additional product with the first product within the product viewing histories of the users; and

representing the set of additional products within a product detail page of the first product to assist users in locating related products during browsing of the online catalog.

33. The method of claim 32, wherein each product viewing history is specific to a particular browsing session of a user.

34. The method of claim 32, wherein representing the set of additional products comprises looking up the products from a periodically generated mapping table in response to a user request for the product detail page.

36. The computer system of claim 35, wherein the first component treats a user access to a detail page of a product as viewing of the product for purposes of generating the product viewing histories.

37. The computer system of claim 35, wherein the product viewing histories are session-specific product viewing histories, such that the second component identifies correlations between product viewing events within the same browsing session.

38. The computer system of claim 35, wherein the second component generates, for each of a plurality of product pairs (product_A, product_B), a value which reflects a frequency with which product_A and product_B were viewed within the same browsing session.

39. The computer system of claim 35, wherein the second component generates a data structure which maps each of a plurality of products to a respective set of related products in which product relationships reflect said correlations between product viewing events.

40. The computer system of claim 39, further comprising a page generation component that uses the data structure to supplement a detail page of a product with a list of related products.

41. The computer system of claim 39, further comprising a third component that uses the data structure to provide real time product recommendations to a user by at least:

identifying multiple products that are of known interest to the user;

for each of the products of known interest, identifying from the data structure a corresponding set of related products;

combining the sets of related products to generate a combined set of related products; and

selecting at least a portion of the combined set of related products to recommend to the user.

42. The computer system of claim 41, wherein the third component identifies the products of known interest to the user by identifying a set of products viewed by the user during a current browsing session.

35. A computer system that determines the relatedness between products of an online collection of products, comprising:

a first component that stores data indicating user-specific product viewing histories of a plurality of users, wherein each product viewing history indicates products viewed by a user during online browsing; and

a second component that processes the data stored by the first component to identify correlations between product viewing events within the product viewing histories, wherein the second component uses the correlations to identify products that are related to one another.