Determining a bid modifier value to maximize a return on investment in a hybrid campaign

• US 9,626,691 B2
• Filed: 05/02/2013
• Issued: 04/18/2017
• Est. Priority Date: 05/02/2013
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Wang, Zhe, Fu, Xiaodong
• Primary Examiner(s): CAO, VINCENT M
• Application Number:
  US13/875,967
• Publication Number:
  US 20150213485A1
• Time in Patent Office: 1,447 Days
• CPC Class Codes: G06Q 30/0242 Determining effectiveness of advertisements
• CPC Class Codes: G06Q 30/0246 Traffic
• CPC Class Codes: G06Q 30/0273 Determination of fees for advertising

Abstract

Systems and methods for determining a bid modifier value include determining, for one or more sets of keywords used in a campaign, a predicted conversion rate for a first device type and a predicted conversion rate for a second device type. Predicted campaign-level conversion rates are determined for the first and second device types using the predicted conversion rates for the one or more sets of keywords used in the campaign. A ratio is also determined between the predicted campaign-level conversion rate for the first device type and the predicted campaign-level conversion rate for the second device type. The ratio may be used as a bid modifier value that relates bids for the first device type to bids for the second device type.

Claims

1. A method of generating a single hybrid campaign formed of separate campaigns to reduce processor and memory consumption during real-time content selection, comprising:

determining, by one or more processors of a content selection service executed by a server, for a first set of keywords used by the content selection service to select one or more content of a first campaign for display via a first device type, a first predicted conversion rate, the first campaign stored in a first memory;

determining, by the content selection service, for a second set of keywords used by the content selection service to select the one or more content of the first campaign for display via the first device type, a second predicted conversion rate;

determining, by the content selection service, for the first set of keywords used by the content selection service to select one or more content of a second campaign for display via a second device type different from the first device type, a third predicted conversion rate, the second campaign stored in a second memory;

determining, by the content selection service, for the second set of keywords used by the content selection service to select the one or more content of the second campaign for display via the second device type, a fourth predicted conversion rate;

calculating, by the content selection service, a first predicted campaign-level conversion rate for the first device type based on a traffic-weighted average of the first predicted conversion rate and the second predicted conversion rate;

calculating, by the content selection service, a second predicted campaign-level conversion rate for the second device type based on a traffic-weighted average of the third predicted conversion rate and the fourth predicted conversion rate;

combining, by the content selection service, the first predicted campaign-level conversion rate for the first device type and the second predicted campaign-level conversion rate for the second device type to determine a bid modifier value;

merging, by the content selection service, the first campaign stored in the first memory with the second campaign stored in the second memory using the determined bid modifier to generate a single hybrid campaign for the first device type and the second device type, the hybrid campaign stored in a third memory; and

using, by the content selection service during real-time content selection, the single hybrid campaign to provide content to at least one of the first device type and the second device type.

2. The method of claim 1, further comprising:

iteratively determining the first predicted conversion rate and the third predicted conversion rate for the first set of keywords and the second predicted conversion rate and the fourth predicted conversion rate for the second set of keywords.

3. The method of claim 2, further comprising:

applying a smoothing function to changes in the bid modifier value responsive to the iterative determinations.

4. The method of claim 1, wherein the traffic-weighted average of the first predicted conversion rate and the third predicted conversion rate is based on a projected number of clicks for the first set of keywords.

5. The method of claim 1, wherein the first device type includes mobile devices and the second device type includes desktop devices.

6. The method of claim 5, wherein the second device type includes tablet devices.

7. A system for generating a single hybrid campaign formed of separate campaigns to reduce processor and memory consumption during real-time content selection, comprising:

a content election service comprising one or more processors configured to;

determine, for a first set of keywords used by the content selection service to select one or more content of a first campaign for display via a first device type, a first predicted conversion rate, the first campaign stored in a first memory;

determine, for a second set of keywords used by the content selection service to select the one or more content of the first campaign for display via the first device type, a second predicted conversion rate;

determine, for the first set of keywords used by the content selection service to select one or more content of a second campaign for display via a second device type different from the first device type, a third predicted conversion rate, the second campaign stored in a second memory;

determine, for the second set of keywords used by the content selection service to select the one or more content of the second campaign for display via the second device type, a fourth predicted conversion rate;

calculate a first predicted campaign-level conversion rate for the first device type based on a traffic-weighted average of the first predicted conversion rate and the second predicted conversion rate;

calculate a second predicted campaign-level conversion rate for the second device type based on a traffic-weighted average of the third predicted conversion rate and the fourth predicted conversion rate;

combine the first predicted campaign-level conversion rate for the first device type and the second predicted campaign-level conversion rate for the second device type to determine a bid modifier value;

merge the first campaign stored in the first memory with the second campaign stored in the second memory using the determined bid modifier to generate a single hybrid campaign for the first device type and the second device type, the hybrid campaign stored in a third memory; and

use, for real-time content selection, the single hybrid campaign to provide content to at least one of the first device type and the second device type.

8. The system of claim 7, wherein the one or more processors are configured to:

iteratively determine the first predicted conversion rate and the third predicted conversion rate for the first set of keywords and the second predicted conversion rate and the fourth predicted conversion rate for the second set of keywords.

9. The system of claim 8, wherein the one or more processors are configured to:

apply a smoothing function to changes in the bid modifier value responsive to the iterative determinations.

10. The system of claim 7, wherein the traffic-weighted average of the first predicted conversion rate and the third predicted conversion rate is based on a projected number of clicks for the first set of keywords.

11. The system of claim 7, wherein the first device type includes mobile devices and the second device type includes desktop devices.

12. The system of claim 11, wherein the second device type includes tablet devices.

13. A computer-readable storage medium having machine instructions stored therein, the instructions being executable by a processor to cause the processor to perform operations to generate a single hybrid campaign formed of separate campaigns to reduce processor and memory consumption during real-time content selection, the operations comprising:

determining, for a first set of keywords used by the content selection service to select one or more content of a first campaign for display via a first device type, a first predicted conversion rate, the first campaign stored in a first memory;

determining, for a second set of keywords used by the content selection service to select the one or more content of the first campaign for display via the first device type, a second predicted conversion rate;

determining, for the first set of keywords used by the content selection service to select one or more content of a second campaign for display via a second device type different from the first device type, a third predicted conversion rate, the second campaign stored in a second memory;

determining, for the second set of keywords used by the content selection service to select the one or more content of the second campaign for display via the second device type, a fourth predicted conversion rate;

calculating a first predicted campaign-level conversion rate for the first device type based on a traffic-weighted average of the first predicted conversion rate and the second predicted conversion rate;

calculating a second predicted campaign-level conversion rate for the second device type based on a traffic-weighted average of the third predicted conversion rate and the fourth predicted conversion rate;

combining, the first predicted campaign-level conversion rate for the first device type and the second predicted campaign-level conversion rate for the second device type to determine a bid modifier value;

merging the first campaign stored in the first memory with the second campaign stored in the second memory using the determined bid modifier to generate a single hybrid campaign for the first device type and the second device type, the hybrid campaign stored in a third memory; and

using, by the content selection service during real-time content selection, the single hybrid campaign to provide content to at least one of the first device type and the second device type.

14. The computer-readable storage medium of claim 13, wherein the operations comprise:

iteratively determining the first predicted conversion rate and the third predicted conversion rate for the first set of keywords and the second predicted conversion rate and the fourth predicted conversion rate for the second set of keywords.

15. The computer-readable storage medium of claim 14, wherein the operations comprise:

applying a smoothing function to changes in the bid modifier value responsive to the iterative determinations.

16. The computer-readable storage medium of claim 13, wherein the traffic-weighted average of the first predicted conversion rate and the third predicted conversion rate is based on a projected number of clicks for the first set of keywords.

17. The computer-readable storage medium of claim 13, wherein the first device type includes mobile devices and the second device type includes desktop devices.