Click modeling for ecommerce

Patent metadata

Field	Value
Publication number	US-9741039-B2
Application number	US-201213680450-A
Country	US
Kind code	B2
Filing date	Nov 19, 2012
Priority date	Nov 22, 2011
Publication date	Aug 22, 2017
Grant date	Aug 22, 2017

Abstract

A system and method of click modeling are disclosed. A ranking algorithm is modified using a click model. The click model makes inferences based on user click data. The user click data includes indications of user intent to transact a shopping action for an item for sale. Item listings are ranked for a search results page using the ranking algorithm in response to receiving a search query. Each item listing comprises an item for sale on an e-commerce site. The user click data can also include user clicks of item listings in a search results page. The shopping action can include bidding on an item for sale or purchasing an item for sale.

First claim

What is claimed is:

1. A system comprising:
at least one processor;
a navigation module, executable by the at least one processor, configured to:
receive a search query from a computing device of a current user;
generate a plurality of item listings for a search results page based on the search query, each item listing of the plurality of item listings comprising an item for sale on an e-commerce site; and
rank the plurality of item listings using a search algorithm; and

a click modeling module, executable by the at least one processor, configured to:
identify historical user click data including indications of user transaction of a shopping action for an item for sale, the shopping action comprising bidding on the item for sale or purchasing the item for sale;
generate a click model based on the historical user click data, the click model configured to determine a probability that a particular user will transact the shopping action for an item listing, the probability being based, at least in part, on a sum of item listings that the particular user had previously examined and had found to be a good deal;
receive historical click data of the current user;
apply the historical click data of the current user to the click model, thereby generating a probability that the current user will transact the shopping action;
modify the ranking of the plurality of item listings based on the generated probability, thereby generating a modified search results page; and
provide the modified search results page to the computing device for display to the current user.

2. The system of claim 1 , wherein the historical user click data further includes user clicks of item listings in a search results page.

3. The system of claim 1 , wherein the click model is further configured to infer a measure of how attractive item listings look in a search results page based on the historical user click data.

4. The system of claim 1 , wherein the click model is further configured to infer a measure of how much of a good shopping deal the item listings are based on the historical user click data.

5. The system of claim 1 , wherein the shopping action is bidding on an item for sale.

6. The system of claim 1 , wherein the shopping action is purchasing an item for sale.

7. A computer-implemented method comprising:
identifying historical user click data including indications of user transaction of a shopping action for an item for sale, the shopping action comprising bidding on the item for sale or purchasing the item for sale;
generating a click model based on the historical user click data, the click model being configured to determine a probability that a particular user will transact the shopping action for an item listing, the probability being based, at least in part, on a sum of item listings that the user has previously examined and has found to be a good deal;
receiving a search query from a computing device of a current user;
generating a plurality of item listings for a search results page based on the search query, each item listing of the plurality of item listings comprising the item for sale on an e-commerce site;
ranking the plurality of item listings using a search algorithm;
receiving historical click data of the current user;
applying the historical click data of the current user to the click model, thereby generating a probability that the current user will transact the shopping action;
modifying the ranking of the plurality of item listings based on the generated probability, thereby generating a modified search results page; and
providing the modified search results page to the computing device for display to the current user.

8. The method of claim 7 , wherein the historical user click data further includes user clicks of item listings in a search results page.

9. The method of claim 7 , wherein the click model is further configured to infer a measure of how attractive item listings look in a search results page based on the historical user click data.

10. The method of claim 7 , wherein the click model is further configured to infer a measure of how much of a good shopping deal the item listings are based on the historical user click data.

11. The method of claim 7 , wherein the shopping action is bidding on an item for sale.

12. The method of claim 7 , wherein the shopping action is purchasing an item for sale.

13. A non-transitory machine-readable storage device storing a set of instructions that, when executed by at least one processor, causes the at least one processor to perform operations comprising:
identifying historical user click data including indications of user transaction of a shopping action for an item for sale, the shopping action comprising bidding on the item for sale or purchasing the item for sale;
generating a click model based on the historical user click data, the click model being configured to determine a probability that a particular user will transact the shopping action for an item listing, the probability being based, at least in part, on a sum of item listings that the user has previously examined and has found to be a good deal;
receiving a search query from a computing device of a current user;
generating a plurality of item listings for a search results page based on the search query, each item listing of the plurality of item listings comprising the item for sale on an e-commerce site;
ranking the plurality of item listings using a search algorithm;
receiving historical click data of the current user;
applying the historical click data of the current user to the click model, thereby generating a probability that the current user will transact the shopping action;
modifying the ranking of the plurality of item listings based on the generated probability, thereby generating a modified search results page; and
providing the modified search results page to the computing device for display to the current user.

14. The device of claim 13 , wherein the historical user click data further includes user clicks of item listings in a search results page.

15. The device of claim 13 , wherein the click model is further configured to infer a measure of how attractive item listings look in a search results page based on the historical user click data.

16. The device of claim 13 , wherein the click model is further configured to infer a measure of how much of a good shopping deal the item listings are based on the historical user click data.

17. The device of claim 13 , wherein the shopping action is bidding on an item for sale.

18. The device of claim 13 , wherein the shopping action is purchasing an item for sale.

Assignees

• Abbas Abdelhalim
• Ebay Inc

Inventors

• Abbas Abdelhalim

Classifications

• G06Q30/00 Primary

  Commerce · CPC title

• G06Q30/0601

  Electronic shopping [e-shopping] · CPC title