Automatic adjustment of advertiser bids to equalize cost-per-conversion among publishers for an advertisement

• US 8,175,914 B1
• Filed: 07/30/2007
• Issued: 05/08/2012
• Est. Priority Date: 07/30/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Benson, Scott S., Tong, Simon
• Primary Examiner(s): Merchant, Shahid R
• Assistant Examiner(s): Goodbody, Joan
• Application Number:
  US11/830,728
• Time in Patent Office: 1,744 Days
• Field of Search: 705/14
• US Class Current: 705/14
• CPC Class Codes: G06Q 10/025 Coordination of plural reservations, e.g. plural trip segments, transportation combined with accommodation
• CPC Class Codes: G06Q 20/105 involving programming of a portable memory device, e.g. IC cards, "electronic purses"
• CPC Class Codes: G06Q 30/0242 Determining effectiveness of advertisements
• CPC Class Codes: G06Q 30/0275 Auctions
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

A learning model is built on a combination of advertiser, publisher and user data. The learning model can be applied to all advertisers in an advertising system. The learning model provides predicted conversion rates for a given advertisement (“ad”) appearing on different publisher networks. A predicted conversion rate represents the probability that a click on a given ad appearing on a given publisher will lead to a conversion. The predicted conversion rates are used to generate a multiplier. The multiplier is used to automatically adjust the advertiser'"'"'s bid (e.g., maximum cost-per-click (CPC)) for the given ad prior to an auction for the ad. Adjusting the advertiser'"'"'s bid equalizes a cost-per-conversion among the publishers for the ad.

Claims

1. A method comprising:

obtaining a first predicted conversion rate for an advertisement appearing on a first publisher property, where the first predicted conversion rate is obtained using a learning model developed from a combination of advertiser and publisher data associated with the advertisement;

obtaining a second predicted conversion rate for the advertisement appearing on a second publisher property;

computing, by one or more processors, a multiplier using the first and second predicted conversion rates, wherein the multiplier is based on a minimum of (i) a first value specifying a cap on the multiplier and (ii) a function of a probability that a conversion will result from a selection of the advertisement by a user on the first publisher property with respect to a probability that a conversion will result from a selection of the advertisement by a user on the second publisher property; and

automatically adjusting a bid of an advertiser associated with the advertisement using the multiplier to equalize a cost-per-conversion for the advertisement between the first and second publisher properties, wherein the cost-per-conversion is an advertising cost to an advertiser to achieve a specified consumer transaction resulting from an advertisement appearing on a publisher property.

2. The method of claim 1, further comprising:

developing the learning model from a combination of advertiser, publisher and user data.

3. The method of claim 1, further comprising:

developing the learning model using a historical predicted conversion rate associated with at least one of the first publisher property or the second publisher property.

4. The method of claim 1, where the advertiser'"'"'s bid is a maximum cost-per-click (CPC).

5. The method of claim 1, further comprising:

obtaining the second predicted conversion rate for the advertisement appearing on the second publisher property using the learning model.

6. The method of claim 4, where automatically adjusting a bid further comprises:

automatically multiplying the CPC by the multiplier.

7. The method of claim 1, wherein obtaining a first predicted conversion rate for an advertisement appearing on a first publisher property, comprises:

determining that a combination of advertiser and publisher data for the advertiser and the first publisher property is below a data confidence threshold;

obtaining a combination of advertiser and publisher data for the advertiser and at least one other publisher property different from the first publisher property; and

obtaining the first predicted conversion rate for the advertisement appearing on the first publisher property based on the combination of advertiser and publisher data for the advertiser and the at least one other publisher property.

8. The method of claim 1, wherein the first value is 1.

9. A system comprising:

a first learning model that provides a first predicted conversion rate for an advertisement appearing on a first publisher property, and provides a second predicted conversion rate for the advertisement appearing on a second publisher property, where the learning model is developed using a combination of advertiser and publisher data associated with the advertisement;

a multiplier generator, implemented by one or more processors, that computes a multiplier using the first and second predicted conversion rates, wherein the multiplier is based on a minimum of (i) a first value specifying a cap on the multiplier and (ii) a function of a probability that a conversion will result from a selection of the advertisement by a user on the first publisher property with respect to a probability that a conversion will result from a selection of the advertisement by a user on the second publisher property; and

a bid adjuster module, implemented by the one or more processors, that automatically adjusts a bid of an advertiser associated with the advertisement using the multiplier generator to equalize a cost-per-conversion for the advertisement between the first and second publisher properties, wherein the cost-per-conversion is an advertising cost to an advertiser to achieve a specified consumer transaction resulting from an advertisement appearing on a publisher property.

10. The system of claim 9, where the learning model is developed from a combination of advertiser, publisher and user data.

11. The system of claim 9, where the learning model is developed from a historical predicted conversion rate associated with at least one of the first publisher property and the second publisher property.

12. The system of claim 9, where the advertiser'"'"'s bid is a maximum cost-per-click (CPC).

13. A non-transitory computer-readable medium having instructions stored thereon, which, when executed by a processor, causes the processor to perform operations comprising:

obtaining a first predicted conversion rate for an advertisement appearing on a first publisher property, where the first predicted conversion rate is obtained using a learning model developed from a combination of advertiser and publisher data associated with the advertisement;

obtaining a second predicted conversion rate for the advertisement appearing on a second publisher property;

computing a multiplier using the first and second predicted conversion rates, wherein the multiplier is based on a minimum of (i) a first value specifying a cap on the multiplier and (ii) a function of a probability that a conversion will result from a selection of the advertisement by a user on the first publisher property with respect to a probability that a conversion will result from a selection of the advertisement by a user on the second publisher property; and

automatically adjusting a bid of an advertiser associated with the advertisement using the multiplier to equalize a cost-per-conversion for the advertisement between the first and second publisher properties, wherein the cost-per-conversion is an advertising cost to an advertiser to achieve a specified consumer transaction resulting from an advertisement appearing on a publisher property.

14. The computer-readable medium of claim 13, where the learning model is developed from a combination of advertiser, publisher and user data.

15. The computer-readable medium of claim 13, where the learning model is developed from a historical predicted conversion rate associated with at least one of the first publisher property and the second publisher property.

16. The computer-readable medium of claim 13, where the advertiser'"'"'s bid is a maximum cost-per-click (CPC).

17. The computer-readable medium of claim 16, where automatically adjusting a bid further comprises:

automatically multiplying the CPC by the multiplier.

18. A system comprising:

a data processing apparatus; and

a data storage apparatus encoded with instructions that when executed by the data processing apparatus cause the data processing apparatus to perform operations comprising;

obtaining a first predicted conversion rate for an advertisement appearing on a first publisher property, where the first predicted conversion rate is obtained using a learning model developed from a combination of advertiser and publisher data associated with the advertisement;

obtaining a second predicted conversion rate for the advertisement appearing on a second publisher property;

computing a multiplier using the first and second predicted conversion rates, wherein the multiplier is based on a minimum of (i) a first value specifying a cap on the multiplier and (ii) a function of a probability that a conversion will result from a selection of the advertisement by a user on the first publisher property with respect to a probability that a conversion will result from a selection of the advertisement by a user on the second publisher property; and

automatically adjusting a bid of an advertiser associated with the advertisement using the multiplier to equalize a cost-per-conversion for the advertisement between the first and second publisher properties, wherein the cost-per-conversion is an advertising cost to an advertiser to achieve a specified consumer transaction resulting from an advertisement appearing on a publisher property.

19. The system of claim 18, wherein the data processing apparatus further performs operations comprising:

developing the learning model from a combination of advertiser, publisher and user data.