QUALITY-WEIGHTED SECOND-PRICE AUCTIONS FOR ADVERTISEMENTS

• US 20140372203A1
• Filed: 06/14/2013
• Published: 12/18/2014
• Est. Priority Date: 06/14/2013
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: POWELL, NATHANIEL EDWARD, JORDAN, PATRICK RICHARD LLOYD, ATHEY, SUSAN C.
• Application Number:
  US13/917,939
• Publication Number:
  US 20140372203A1
• Time in Patent Office: Days
• US Class Current: 705/14.42
• CPC Class Codes: G06Q 30/0243 Comparative campaigns
• CPC Class Codes: G06Q 30/0246 Traffic
• CPC Class Codes: G06Q 30/0275 Auctions

Abstract

Methods, computer systems, and computer-storage media are provided for utilizing a quality-weighted generalized second-price auction to rank, price, and allocate advertisements for presentation on a results page. Advertisements responsive to a search query are identified, and a rank score is determined for each of the advertisements. The rank score is determined based on a quality score associated with each of the advertisements as well as a bid price for each of the advertisements. The satisfied click-through-rate for an advertisement is used as a proxy metric for the quality score. The quality score is further used to determine the cost-per-click of the advertisement. Certain quality metrics may also be utilized to determine the number of advertisements that should be presented on the results page in order to maximize publisher revenue and minimize user cost.

Claims

1. One or more computer-storage media having computer-executable instructions embodied thereon that, when executed, perform a method of ranking advertisements that are presented on a results page using quality weighting in a generalized second-price auction, the method comprising:

identifying a set of advertisements that is relevant to a search query;

determining a rank score for each advertisement in the set of advertisements, the rank score determined based on a quality score of the each advertisement and a bid price of the each advertisement, wherein the quality score is based in part on a satisfied click-through-rate for the each advertisement;

ranking the set of advertisements based on the rank scores; and

communicating a plurality of top-ranked advertisements of the set of advertisements to a publisher who presents in ranked order the plurality of top-ranked advertisements on the results page.

2. The media of claim 1, wherein the set of advertisements includes advertisements that are exact matches to the search query and advertisements that are broad matches to the search query.

3. The media of claim 1, wherein the rank score is determined using the following formula:

Rank Score_i=q_i*b_i

where Rank Score, is the rank score for the i^th advertisement,

q_i is the quality score for the i^th advertisement, and

b_i is the bid price of the i^th advertisement.

4. The media of claim 1, wherein the quality score of the each advertisement comprises a value between 0 and 1, and wherein the quality score provides an indication of a proportion of users whose needs are met by the each advertisement and the each advertisement'"'"'s associated landing page(s).

5. The media of claim 1, wherein the satisfied click-through-rate comprises a proportion of users who selected the each advertisement and dwelled on the each advertisement'"'"'s associated landing page for a predetermined period of time.

6. The media of claim 5, wherein the predetermined period of time is 20 seconds or greater.

7. The media of claim 5, wherein the predetermined period of time is 30 seconds or greater.

8. The media of claim 1, wherein the quality score of the each advertisement is further based in part on one or more of the following:

a proportion of users who access a secure page after selecting the each advertisement,

a proportion of users who terminate an ongoing search session after selecting the each advertisement, or

a proportion of users who make a purchase after selecting the each advertisement.

9. The media of claim 1, wherein the results page comprises a search engine results page.

10. One or more computer-storage media having computer-executable instructions embodied thereon that, when executed, perform a method of ranking and pricing advertisements that are presented on a results page using quality weighting in a generalized second-price auction, the method comprising:

identifying a set of advertisements that is relevant to a search query;

determining a rank score for each advertisement in the set of advertisements, the rank score determined based on a quality score of the each advertisement and a bid price of the each advertisement, wherein the quality score is based in part on a satisfied click-through-rate for the each advertisement;

ranking the set of advertisements based on the rank scores; and

determining a cost-per-click (CPC) for the each advertisement in the set of advertisements, wherein for a given advertisement in the set of advertisements;

when the given advertisement is an exact match to the search query, determining the CPC based on a quality score of an advertisement ranked directly below the given advertisement, a bid price of the advertisement ranked directly below the given advertisement, and a quality score of the given advertisement, and

when the given advertisement is a broad match to the search query, determining the CPC based on a probability that an advertisement ranked directly below the given advertisement will receive a click, a probability that a user'"'"'s need is met given there is a click on the advertisement ranked directly below the given advertisement, a bid price of the advertisement ranked directly below the given advertisement, and a probability that the given advertisement will receive a click.

11. The media of claim 10, wherein determining the CPC for the given advertisement that is the exact match to the search query comprises using the following formula:

12. The media of claim 10, wherein determining the CPC for the given advertisement that is the broad match to the search query comprises using the following formula:

13. One or more computer-storage media having computer-executable instructions embodied thereon that, when executed, perform a method of determining a number of advertisements to be presented in an advertisement block on a results page in a generalized second-price auction, the method comprising:

defining an expected satisfied revenue when n top-ranked advertisements are presented in the advertisement block;

defining a user cost when n top-ranked advertisements are presented in the advertisement block;

defining an objective function that generates a solution based on the expected satisfied revenue and the user cost for each n top-ranked advertisements in the advertisement block;

for each value of n, using the objective function to generate a solution;

determining the n value associated with the solution having the highest value, wherein the n value associated with the solution having the highest value is equal to the number of advertisements to be presented in the advertisement block on the results page such that expect satisfied revenue is maximized and user cost is minimized; and

communicating the determined number of advertisements to a publisher who presents the determined number of advertisements in the advertisement block on the results page.

14. The media of claim 13, wherein the expected satisfied revenue is based in part on a satisfied click-through-rate for each of the n top-ranked advertisements.

15. The media of claim 13, wherein the user cost is based on vertical space on a results page that is occupied by the n top-ranked advertisements.

16. The media of claim 15, wherein the user cost is further based on a probability of a satisfied click on the n top-ranked advertisements compared to a probability of a satisfied click on the results page as a whole.

17. The media of claim 16, wherein the user cost is inversely proportional to the probability of the satisfied click on the n top-ranked advertisements compared to the probability of the satisfied click on the results page as a whole.

18. The media of claim 13, further comprising:

subsequent to determining the number of advertisements to be presented in the advertisement block, determining a minimum reserve price for the lowest-ranked advertisement that is presented in the advertisement block, wherein the minimum reserve price comprises a maximum of the following;

a predetermined minimum reserve price,

a reserve price based in part on the user cost, or

a bid price of an advertisement that is ranked directly below the lowest-ranked advertisement that is presented on the results page.

19. The media of claim 18, wherein the reserve price based in part on the user cost is generated by equating the objective function that utilizes the n value equal to the number of advertisement presented in the advertisement block to an objective function that utilizes a k value equal to a number of advertisements that generates the next-highest-valued solution.

20. The media of claim 18, wherein the predetermined minimum reserve price is set by the publisher.