Keyword-based content suggestions

• US 8,463,830 B2
• Filed: 01/05/2007
• Issued: 06/11/2013
• Est. Priority Date: 01/05/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Criou, Nathalie, Mahadevan, Chandramouli, Venkatakrishnan, Sridhar
• Primary Examiner(s): Chen, Susan
• Application Number:
  US11/620,421
• Publication Number:
  US 20080168032A1
• Time in Patent Office: 2,349 Days
• Field of Search: 707 3- 5, 707/771, 709/203, 709/204, 709/223, 709/224, 709/217, 705/26, 705/27, 705/37
• US Class Current: 707/899
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/0256 User search

Abstract

A system and related methods suggest content based on user input and another metric. In one implementation, web-sites are suggested to advertisers in response to keyword input and by factoring in how often such web-sites were selected or “clicked on” as a result of corresponding search queries. Search logs are processed to determine how often certain query terms led to web-sites being selected. Web-sites are ranked accordingly. Keywords from advertisers are matched to the web-site rankings to present the top web-sites.

Claims

1. A computer-implemented method comprising:

identifying, by a computer system, at least one advertiser keyword that an advertiser has associated with an advertisement;

in response to identifying the at least one advertiser keyword;

identifying, by the computer system and from a plurality of user submitted search queries, a set of matching search queries that match the at least one advertiser keyword;

for each of one or more user search queries in the set of matching search queries;

identifying a set of web-sites that were identified by search results previously provided for the user search query; and

determining, for each of two or more web-sites in the set of identified web-sites, a total number of users that clicked on the search results that identified the web-site;

ranking, by the computer system, each of the two or more web-sites in the set of identified web-sites based on the total number of users that clicked on the search results for the web-site;

selecting, by the computer system, at least one of the two or more web-sites as a suggested web-site for which presentation of the advertisement on the web-site is suggested, the suggested web-site being selected based on the ranking of the two or more web-sites; and

providing, by the computer system and to the advertiser, a list identifying the suggested web-site.

2. The method of claim 1, wherein:

identifying a set of web-sites comprises;

creating, by the computer system, a keyword content vector for the at least one keyword, the keyword content vector including phrases that are included in the at least one keyword;

creating, by the computer system, content vectors for two or more web-sites, the content vector for each web-site including phrases that were included in the search queries that preceded user clicks on the search results for the web-site; and

identifying, as the set of web-sites, the two or more web-sites having content vectors that match the keyword content vector.

3. The method of claim 1, further comprising receiving, from the advertiser, a selection of at least one suggested web-site from the list, the selection by the advertiser indicating that the advertisement is available for presentation with the selected web-site in response to a user request for presentation of the selected web-site.

4. The method of claim 1, wherein selecting a set of suggested web-sites comprises selecting the set of web-sites having at least a minimum ranking.

5. A computer implemented method of targeting an on-line advertisement, comprising:

receiving, by a computer system and from the advertiser, a keyword for an advertisement, the keyword being one or more words;

in response to receiving the keyword from the advertiser;

determining, by the computer system, a content vector for the keyword, the content vector including at least one word from the keyword;

identifying, by the computer system, data specifying a search query that matches the content vector, data specifying a plurality of search queries that were received from users, and data specifying whether the users selected search results that were provided, based on the search queries, to the users;

identifying, by the computer system and based on the identified data, web-sites that have been requested by the users through the users'"'"' clicks on the search results and a total number of times that each of the web-sites has been requested through the users'"'"' clicks on the search results;

selecting, by the computer system and for the advertiser, a set of suggested web-sites for which presentation of the advertisement on the web-site is suggested;

providing, by the computer system and to the advertiser, information identifying a list of suggested web-sites, from the set of suggested web-sites, for targeting presentation of the advertisement, the list of suggested web-sites being a list of the suggested web-sites on which presentation of the advertisement is suggested, each suggested web-site being included in the list of suggested web-sites based on a total number of times that the suggested web-site was requested through the users'"'"' clicks on the search results; and

enabling selection, by the advertiser and from the list of suggested web-sites, of one or more of the suggested web-sites as web-sites on which the advertisement will be presented.

6. The method of claim 5, further comprising:

accessing a data structure that includes data indicating a number of times the content vector occurred in previous queries that resulted in selections of the web-sites; and

ranking the web-sites according to the number of times each web-site was selected in response to a query in which the content vector occurred.

7. The method of claim 6, wherein a threshold number of ranked web-sites are presented based on the number of times the web-sites were selected in response to a query in which the content vector occurred.

8. The method claim 5, further comprising determining a set of content vectors corresponding to the at least one keyword, wherein providing information to the advertiser identifying the suggested web-sites includes identifying the suggested web-sites based on the set of content vectors for the at least one keyword.

9. A computer-implemented method, comprising:

receiving user search queries for which search results for web-sites were provided to users and data specifying a number of the users that clicked on the search results;

creating, by one or more computers and based on the data, first content vectors for the web-sites, each first content vector for each website including phrases from the search queries that were submitted by the users that clicked on the search results for the web-site, the first content vector for each web-site further specifying a total number of users that clicked on the search results for the web-site, the creating comprising;

for at least one web-site;

creating a two-word phrase that is included in the search queries that were submitted by the users that clicked on the search results for the at least one web-site;

creating two or more one-word phrases that are included in the search queries that were submitted by the users that clicked on the search results for the web-site; and

creating a first content vector that includes the two-word phrase and the two or more one-word phrases, wherein creating the first content vector that includes the two-word phrase and the two or more one-word phrases comprises assigning weights to the two-word phrase and the two or more one-word phrases, the assignment of weights being based on total numbers of clicks on the search results for the at least one web-site when the search results were provided in response to phrases that are included in the first content vector;

receiving, from an advertiser, two or more phrases;

creating, by the one or more computers, a second content vector for the two or more phrases, comprising;

creating one or more two-word phrases that are included in the two or more phrases;

creating two or more one-word phrases that are included in the two or more phrases;

creating a second content vector that includes the one or more two word phrases and the two or more one-word phrases; and

assigning a weight to each phrase that is included in the second content vector, the weight assigned to each respective phrase being based on a number of times that the advertiser has used the respective phrase to target advertisements;

selecting, by the one or more computers, a set of the web-sites having the first content vectors that match the second content vector; and

providing, by the one or more computers, a list of suggested web-sites to the advertiser, the list of suggested web-sites specifying one or more web-sites in the set based on the total number of users that clicked on the search results for the web-sites in the set.