Systems and methods for predicting advertising revenue

• US 7,835,939 B1
• Filed: 10/31/2007
• Issued: 11/16/2010
• Est. Priority Date: 10/31/2007
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: AOL Advertising Incorporated (Apollo Global Management, Inc.)
• Inventors: Karlsson, Niklas
• Primary Examiner(s): DIXON, THOMAS A
• Application Number:
  US11/979,098
• Time in Patent Office: 1,112 Days
• Field of Search: 705/10
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

Methods and system are provided to predict data related to advertising. Predicted data may include impressions, clicks, actions, or other data related to the advertisement, such as the ratio of impressions to clicks or the ratio of clicks to actions. Methods for predicting revenue of an advertising campaign use a revenue predictor in a campaign control computer while the campaign is ongoing. Based on a starting point in time during the advertising campaign when a starting revenue and volume data through the starting point in time were known and a target point in time, a cumulative predicted revenue can be determined. Based on an estimated revenue generated by the advertising campaign from a time just after the starting point in time until a selected point in time and a predicted revenue from a time just after the selected point in time until the target point in time is computed, the cumulative predicted revenue for the advertising campaign through the target point in time is determined.

Claims

1. A computer-implemented method for predicting revenue of an advertising campaign using a revenue predictor in a campaign control computer while the campaign is ongoing, comprising:

receiving, by a controller of the campaign control computer, a starting point in time, during the advertising campaign, when a starting revenue and volume data through the starting point in time were known;

receiving, by the controller, a target point in time, at which a cumulative predicted revenue is to be determined representing the cumulative revenue generated by the advertising campaign up to the target point in time;

dividing, by the controller, the time between the starting point in time and the target point in time into at least two intervals based on a selected point in time at which less than all of the starting revenue and the volume data were known;

computing, by the controller, an estimated revenue generated by the advertising campaign from a time just after the starting point in time until the target point in time;

computing, by the controller, a predicted revenue generated by the advertising campaign from a time just after the selected point in time until the target point in time; and

determining, by the controller, the cumulative predicted revenue for the advertising campaign through the target point in time based on the starting revenue, the estimated revenue, and the predicted revenue.

2. The computer-implemented method of claim 1, wherein computing, by the controller, the estimated revenue further includes:

computing, by the controller, the estimated revenue using a known impression volume for the advertising campaign through the selected point of time and an estimated impression-to-click probability for the advertising campaign.

3. The computer-implemented method of claim 1, wherein computing, by the controller, the predicted revenue further includes:

computing, by the controller, the predicted revenue using volume models.

4. The computer-implemented method of claim 1, wherein the volume data includes at least one of an impression volume, a click volume, and an action volume.

5. The computer-implemented method of claim 1, wherein the starting revenue and the volume data through the starting point are considered to be known when they can be estimated to a predetermined accuracy.

6. The computer-implemented method of claim 1, further comprising:

determining, by the controller, a control signal based on the cumulative predicted revenue; and

applying the control signal, by the campaign control computer, to adjust behavior of the advertising campaign.

7. The computer-implemented method of claim 1, further comprising:

determining, by the controller, a control signal based on the cumulative predicted revenue; and

applying the control signal, by the campaign control computer, to alter placement of an advertisement during the advertising campaign.

8. A computer-implemented method for predicting revenue of an advertising campaign using a revenue predictor in a campaign control computer while the campaign is ongoing, comprising:

receiving, by a controller of the campaign control computer, a starting point in time, during the advertising campaign, when a starting revenue, click volume data, impression volume data, and conversion volume data through the starting point in time were known;

determining, by the controller, an estimated revenue for a portion of the advertising campaign from a time just after the starting point in time until a second point in time when the click volume data, the impression volume data, and the conversion volume data were known but the starting revenue was not known;

computing, by the controller, a probability-predicted revenue for a portion of the advertising campaign from a time just after the second point in time until a third point in time when the click volume data and the impression volume data were known but the conversion volume data and the starting revenue were not known;

computing, by the controller, a model-predicted revenue generated by the advertising campaign from a time just after the third point in time until a target point in time when click the volume data, the impression volume data, the conversion volume data, and the starting revenue are not known; and

estimating, by the controller, a cumulative revenue for the advertising campaign through the target point in time by combining the starting revenue, the estimated revenue, the probability-predicted revenue, and the model-predicted revenue.

9. The computer-implemented method of claim 8, wherein estimating, by the controller, the cumulative revenue further includes:

estimating, by the controller, the cumulative revenue using an impression-to-click probability for the advertising campaign.

10. The computer-implemented method of claim 8, wherein computing, by the controller, the model-predicted revenue further includes:

computing, by the controller, the model-predicted revenue using volume models to predict the click volume data, the impression volume data, and the conversion volume data.

11. The computer-implemented method of claim 8, wherein data is considered to be known when it can be estimated to a predetermined accuracy.

12. The computer-implemented method of claim 8, further comprising:

determining, by the controller, a control signal based on the cumulative predicted revenue; and

applying the control signal, by the campaign control computer, to adjust behavior of the advertising campaign.

13. The computer-implemented method of claim 8, further comprising:

determining, by the controller, a control signal based on the cumulative predicted revenue; and

applying the control signal, by the campaign control computer, to alter placement of an advertisement during the advertising campaign.

14. A system for predicting revenue of an advertising campaign while the campaign is ongoing, comprising:

a campaign control computer; and

a revenue predictor configured to execute a computer-implemented method, the computer-implemented method including;

receiving a starting point in time, during the advertising campaign, when a starting revenue and volume data through the starting point in time were known;

receiving a target point in time, at which a cumulative predicted revenue is to be determined representing a cumulative revenue generated by the advertising campaign up to the target point in time;

dividing the time between the starting point and the target point into at least two intervals based on a selected point in time at which less than all of the cumulative revenue and the volume data were known;

computing an estimated revenue generated by the advertising campaign from a time just after the starting point in time until the selected point in time;

computing a predicted revenue generated by the advertising campaign from a time just after the selected point in time until the target point in time; and

determining the cumulative predicted revenue for the advertising campaign through the target point in time based on the starting revenue, the estimated revenue, and the predicted revenue.

15. The system of claim 14, wherein the computer-implemented method executed by the revenue predictor further includes:

computing the estimated revenue using a known impression volume for the advertising campaign through the selected point of time and an estimated impression-to-click probability for the advertising campaign.

16. The system of claim 14, wherein the computer-implemented method executed by the revenue predictor further includes:

determining the predicted revenue using volume models.

17. The system of claim 14, wherein the volume data includes at least one of an impression volume, a click volume, and an action volume.

18. The system of claim 14, wherein the starting revenue and the volume data through the starting point are considered to be known when they can be estimated to a predetermined accuracy.

19. The system of claim 14, wherein the campaign control computer is further configured to:

determine a control signal based on the cumulative predicted revenue; and

apply the control signal to adjust the behavior of the advertising campaign.

20. The system of claim 14, wherein the campaign control computer is further configured to:

determine a control signal based on the cumulative predicted revenue; and

apply the control signal to alter placement of an advertisement during the advertising campaign.

21. A computer-implemented method for predicting revenue of an advertising campaign using a revenue predictor in a campaign control computer while the campaign is ongoing, comprising:

receiving, by a controller of the campaign control computer, a starting point in time, during the advertising campaign, when a starting revenue and volume data through the starting point in time were known;

receiving, by the controller, a target point in time, at which a cumulative predicted revenue is to be determined representing the cumulative revenue generated by the advertising campaign up to the target point in time;

dividing, by the controller, the time between the starting point in time and the target point in time into at least two intervals based on a selected point in time at which less than all of the starting revenue and the volume data were known;

computing, by the controller, an estimated revenue generated by the advertising campaign from a time just after the starting point in time until the selected point in time;

computing, by the controller, a predicted revenue generated by the advertising campaign from a time just after the selected point in time until the target point in time;

determining, by the controller, the cumulative predicted revenue for the advertising campaign through the target point in time based on the starting revenue, the estimated revenue, and the predicted revenue; and

altering, by the controller, a placement of an advertisement during the advertising campaign based on the cumulative predicted revenue.

22. The computer-implemented method of claim 21, wherein computing, by the controller, the estimated revenue further includes:

computing, by the controller, the estimated revenue using a known impression volume for the advertising campaign through the selected point of time and an estimated impression-to-click probability for the advertising campaign.

23. The computer-implemented method of claim 21, wherein computing, by the controller, the predicted revenue further includes:

computing, by the controller, the predicted revenue using volume models.

24. The computer-implemented method of claim 21, wherein the volume data includes at least one of an impression volume, a click volume, and an action volume.

25. A computer-implemented method for predicting revenue of an advertising campaign using a revenue predictor in a campaign control computer while the campaign is ongoing, comprising:

receiving, by a controller of the campaign control computer, a starting point in time, during the advertising campaign, when a starting revenue and volume data through the starting point in time were known;

receiving, by the controller, a first target point in time, at which a first cumulative predicted revenue is to be determined representing the cumulative revenue generated by the advertising campaign up to the target point in time;

receiving, by the controller, a second target point in time, at which a second cumulative predicted revenue is to be determined representing the cumulative revenue generated by the advertising campaign up to the target point in time;

computing, by the controller, a first estimated revenue generated by the advertising campaign from a first time just after the starting point in time until the first target point in time;

computing, by the controller, a second estimated revenue generated by the advertising campaign from a second time just after the starting point in time until the second target point in time;

computing, by the controller, a first predicted revenue generated by the advertising campaign from the first time just after the selected point in time until the first target point in time;

computing, by the controller, a predicted revenue generated by the advertising campaign from a second time just after the selected point in time until the second target point in time;

determining, by the controller, the first cumulative predicted revenue for the advertising campaign through the first target point in time based on the starting revenue, the first estimated revenue, and the first predicted revenue;

determining, by the controller, the second cumulative predicted revenue for the advertising campaign through the second target point in time based on the starting revenue, the second estimated revenue, and the second predicted revenue;

altering, by the controller, a placement of a first advertisement of the advertising campaign based on the first cumulative predicted revenue; and

altering, by the controller, a placement of a second advertisement of the advertising campaign based on the second cumulative predicted revenue.

26. The computer-implemented method of claim 25, wherein the volume data includes at least one of an impression volume, a click volume, and an action volume.