Generation and contextual presentation of statistical data reflective of user selections from an electronic catalog

• US 8,326,658 B1
• Filed: 06/09/2004
• Issued: 12/04/2012
• Est. Priority Date: 04/12/2004
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Lee, Dennis, Henne, Randal, Holden, Jeffrey, Smith, Brent, Vadon, Eric, Hsu, David
• Primary Examiner(s): Jasmin, Lynda
• Assistant Examiner(s): SANTIAGO, LUIS F
• Application Number:
  US10/864,288
• Time in Patent Office: 3,100 Days
• Field of Search: 705/10, 705/26, 705/37, 705/7.11, 705/26.1, 707/3
• US Class Current: 705/7.11
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0603 Catalogue ordering

Abstract

An electronic catalog system monitors browsing actions of users to generate user activity data. A data mining component analyzes the user activity data to generate conditional probability data that is presented to users contextually to assist users in selecting items from the electronic catalog. The conditional probability data reflects the frequencies with which users that have performed specific actions, such as viewing a particular item, set of items or web page, or conducting a specific search, have thereafter acquired specific items from the catalog. For example, in one embodiment, when a user selects a first item for viewing in the catalog, the user is notified, for at least one second item, of a frequency with which users who have viewed the first item and then made a purchase have acquired the second item.

Claims

1. A data mining method, comprising:

programmatically analyzing user activity data that reflects item viewing events and item acquisition events to calculate, for a pair of items item_A, item_B, a frequency with which users that have acquired an item after viewing item_A have acquired item_B; and

when a user selects item_A for viewing in an electronic catalog of items, notifying the user of said frequency with which users that have acquired an item after viewing item_A have acquired item_B, to assist the user in selecting an item to acquire;

wherein the method is performed automatically by a system that comprises computer hardware.

2. The data mining method of claim 1, wherein a decision to notify the user of said frequency is responsive to a programmatic determination that item_A and item_B are likely substitutes by virtue of a relatively high frequency with which users who viewed item_A during a browsing session also viewed item_B during the browsing session.

3. The data mining method of claim 1, wherein the frequency is calculated without taking into consideration at least one of the following:

(a) items acquired outside an item category in which item_A falls, (b) acquisition events that occurred more than a selected time period after viewing item_A, (c) acquisition events that occurred within browsing sessions in which item A was not viewed.

4. The data mining method of claim 1, further comprising calculating, and notifying the user of, a frequency with which users that have acquired an item after viewing item_A have acquired item_A, to assist the user in selecting between at least item_A and item_B.

5. The data mining method of claim 4, wherein the user is notified of said frequencies via messaging included on an item detail page for item_A.

6. The data mining method of claim 1, wherein, for purposes of calculating the frequency, a user who viewed item_A and acquired item_B within a selected time interval is treated as having acquired item_B after viewing item_A.

7. The data mining method of claim 1, wherein the user is notified of the frequency by notifying the user of a percentage of users that acquired item_B.

13. The data mining method of claim 1, wherein the item acquisition events are item purchase events.

8. A computer system, comprising:

a monitoring component that monitors user actions performed with respect to acquirable items in an electronic catalog to generate user activity data reflective of actions of a plurality of users;

a data mining component that analyzes the user activity data to determine, for a selected pair of items item_A, item_B in the electronic catalog, what percentage of a set of users who viewed item_A and acquired an item acquired item_B; and

a server that is responsive to selection of item_A by a user for viewing in the electronic catalog by notifying the user of said percentage of users that acquired item_B, said server comprising computer hardware.

9. The computer system of claim 8, wherein the data mining component determines the percentage without taking into consideration at least one of the following:

(a) acquisitions of items falling outside an item category of item_A, (b) item acquisition events that occurred more than a selected time period after corresponding viewing events of item_A, (c) item acquisition events that occurred within browsing sessions in which item_A was not viewed.

10. The computer system of claim 8, wherein the data mining component determines the percentage without taking into consideration item acquisition events that occurred prior to corresponding viewing events of item_A.

11. The computer system of claim 8, wherein the data mining component additionally calculates, and the server notifies the user of, a percentage of the set of users that acquired item_A.

12. The computer system of claim 8, wherein the server notifies the user of said percentage by serving an item detail page for item_A with a representation of said percentage included in said item detail page.

14. The computer system of claim 8, wherein the data mining component calculates said percentage based on purchases made by users who have viewed item A.