System identification, estimation, and prediction of advertising-related data

• US 7,827,128 B1
• Filed: 06/04/2008
• Issued: 11/02/2010
• Est. Priority Date: 05/11/2007
• Status: Active Grant

• Current Assignee: Mercury Kingdom Assets Limited
• Original Assignee: AOL Advertising Incorporated (Apollo Global Management, Inc.)
• Inventors: Batiz-Mendoza, Jaime, Karlsson, Niklas, Zhang, Jianlong
• Primary Examiner(s): HOLMES, MICHAEL B
• Application Number:
  US12/155,442
• Time in Patent Office: 881 Days
• Field of Search: 706/21
• US Class Current: 706/21
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0241 Advertisements
• CPC Class Codes: G06Q 30/0242 Determining effectiveness of advertisements
• CPC Class Codes: G06Q 30/0243 Comparative campaigns
• CPC Class Codes: G06Q 30/0246 Traffic

Abstract

In accordance with the invention, a system, method, and apparatus for analyzing advertisement-related data are presented, which may include receiving data related to an aspect of an advertisement and modeling the aspect of the advertisement with a mathematical model. The mathematical model may include a control-signal-related component, a control-signal-independent component, and an error component. Each component may be updated based on at least one of a control signal, the received data, and a previous state of at least one of the components. An updated model may be created based on the updated components. The system, method, and apparatus may also include predicting the aspect of the advertisement using the updated model. Exemplary aspects of and data related to the advertisement may include one or more of the following: a number of impressions, “clicks,” or “conversions” and/or the impression-to-conversion, impression-to-click, or click-to-conversion ratios.

Claims

1. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

modeling an aspect of an online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component;

receiving data related to the aspect of the online advertisement;

updating each component of the mathematical model based on at least one of a control signal related to the received data, the received data, and a previous state of at least one of the components, to create an updated model of the aspect of the online advertisement; and

predicting a future state of the aspect of the online advertisement using the updated model.

2. The method of claim 1, further comprising the step of modifying an allocation of the online advertisement by adjusting the control signal related to the received data.

3. The method of claim 1, further comprising including an error component in the mathematical model, wherein the error component is a multiplicative error component.

4. The method of claim 1, wherein the received data comprises a number of times that the online advertisement has been shown.

5. The method of claim 1, wherein the received data comprises a number of times that the online advertisement has been selected by a user.

6. The method of claim 1, wherein the received data comprises a number of times that a conversion has occurred based on the online advertisement.

7. The method of claim 6, wherein the conversion based on the online advertisement is a sale of an item.

8. The method of claim 6, wherein the conversion based on the online advertisement is selected from the group consisting of filling out a form, taking a survey, and watching a sequence of web pages.

9. The method of claim 1, wherein the received data comprises a probability that an impression will result in a selection.

10. The method of claim 1, wherein the received data comprises a probability that an impression will result in a conversion.

11. The method of claim 1, wherein the received data comprises a probability that a selection by a user will result in a conversion.

12. The method of claim 11, wherein the probability that a selection by a user will result in a conversion is a probability that selection by the user will result in a sale.

13. The method of claim 1, wherein the online advertisement is an advertisement displayed on a web page of the Internet.

14. The method of claim 1, wherein the online advertisement is an advertisement displayed on an Internet-enabled television.

15. The method of claim 1, wherein the online advertisement is displayed on an Internet-enabled cell phone or personal digital assistant.

16. The method of claim 1, wherein the online advertisement is one of a plurality of online advertisements and wherein the method further comprises:

choosing which online advertisement among the plurality of online advertisements to display based at least in part on the prediction of the aspect of the online advertisement.

17. The method of claim 16, wherein the choice of which online advertisement to display is further based on the control signal.

18. The method of claim 1, wherein the mathematical model is a log-linear model of the form

19. The method of claim 1, wherein the mathematical model is a logit-linear model of the form

20. The method of claim 1, wherein the mathematical model is updated using an adaptive estimation scheme.

21. The method of claim 1, wherein the mathematical model is updated using a standard Kalman filter.

22. The method of claim 1, wherein the mathematical model is updated using an extended Kalman filter.

23. The method of claim 1, wherein the mathematical model is updated using an unscented Kalman filter.

24. The method of claim 3, wherein the error component comprises an autoregressive moving average process.

25. The method of claim 1, wherein the control-signal-independent, time-based periodic component comprises a periodic function.

26. The method of claim 25, wherein the periodic function is a sum of one or more sine or cosine functions of periods that are multiples of 24 hours or 168 hours.

27. The method of claim 1, wherein the aspect is selected from the group consisting of:

a number of times that the online advertisement is shown;

a number of times that the online advertisement is selected by a user;

a number of times that a conversion occurs based on the online advertisement;

a probability that an impression results in selection by a user;

a probability that an impression results in a sale; and

a probability that a selection by a user results in a conversion.

28. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

receiving data related to an aspect of an online advertisement;

modeling the aspect of the online advertisement with a mathematical model, comprising a control-signal-related component and a control-signal-independent, time-based periodic component;

updating the control-signal-related component of the mathematical model based on at least one of a control signal related to the received data, the received data, and a previous state of the control-signal-related component;

updating the control-signal-independent, time-based periodic component of the mathematical model based on at least one of the received data and a previous state of the control-signal-independent, time-based periodic component;

generating an updated model of the aspect of the online advertisement based on the updated control-signal-related component and the updated control-signal-independent, time-based periodic component; and

predicting a future state of the aspect of the online advertisement using the updated model.

29. The method of claim 28, wherein the mathematical model further comprises an error component.

30. The method of claim 29, wherein the error component comprises an autoregressive moving average process.

31. The method of claim 28, wherein the online advertisement is an advertisement displayed on a web page of the Internet.

32. The method of claim 28, wherein the online advertisement is one of a plurality of online advertisements and wherein the method further comprises:

choosing which online advertisement among the plurality of online advertisements to display based at least in part on the prediction of the future state of the aspect of the online advertisement.

33. The method of claim 28, wherein the received data relates to an effect that a change to the control signal has on the aspect.

34. The method of claim 28, wherein the mathematical model is a log-linear model of the form

35. The method of claim 28, wherein the mathematical model is a logit-linear model of the form

36. The method of claim 28, wherein the mathematical model is updated using an adaptive estimation scheme.

37. The method of claim 28, wherein the mathematical model is updated using a standard Kalman filter.

38. The method of claim 28, wherein the mathematical model is updated using an extended Kalman filter.

39. The method of claim 28, wherein the mathematical model is updated using an unscented Kalman filter.

40. The method of claim 28, wherein the control-signal-related component and the control-signal-independent, time-based periodic component comprise periodic functions.

41. The method of claim 28, wherein the aspect is selected from the group consisting of:

a number of times that the online advertisement is shown;

a number of times that the online advertisement is selected by a user;

a number of times that a conversion occurs based on the online advertisement;

a probability that an impression results in selection by a user;

a probability that an impression results in a sale; and

a probability that a selection by a user results in a conversion.

42. A system for analyzing online advertisement-related data, the system comprising:

a plant estimation module that;

receives data related to an aspect of an online advertisement;

models the aspect of the online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component; and

updates each component of the mathematical model based on at least one of a control signal related to the received data, the received data, and a previous state of at least one of the components, to create an updated model; and

a control module that predicts a future state of the aspect of the online advertisement using the updated model.

54. The system of claim 42, wherein the control module is further configured to modify an allocation of the online advertisement by adjusting the control signal related to the received data.

43. A system for analyzing online advertisement-related data, the system comprising:

a plant estimation module that;

receives data related to an aspect of an online advertisement;

models the aspect of the online advertisement with a mathematical model, comprising a control-signal-related component and a control-signal-independent, time-based periodic component;

updates the control-signal-related component of the mathematical model based on at least one of a control signal related to the received data, the received data, and a previous state of the control-signal-related component;

updates the control-signal-independent, time-based periodic component of the mathematical model based on at least one of the received data and a previous state of the control-signal-independent, time-based periodic component; and

generates an updated model of the aspect of the online advertisement based on the updated control-signal-related component and the updated control-signal-independent, time-based periodic component; and

a control module that predicts a future state of the aspect of the online advertisement using the updated model.

55. The system of claim 43, wherein the control module is further configured to modify an allocation of the online advertisement by adjusting the control signal related to the received data.

44. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

receiving data related to an aspect of an online advertisement;

modeling the aspect of the online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component;

updating each component of the mathematical model based on at least one of a control signal, the received data, and a previous state of at least one of the components, to create an updated model; and

predicting a future state of the aspect of the online advertisement using the updated model,

wherein the received data comprises a probability that an impression of the online advertisement will result in a conversion.

45. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

receiving data related to an aspect of an online advertisement;

modeling the aspect of the online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component;

updating each component of the mathematical model based on at least one of a control signal, the received data, and a previous state of at least one of the components, to create an updated model; and

predicting a future state of the aspect of the online advertisement using the updated model,

wherein the received data comprises a probability that a selection of the online advertisement by a user will result in a conversion.

46. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

modeling an aspect of an online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component;

receiving data related to the aspect of the online advertisement;

predicting a future state of the aspect of the online advertisement, based on the mathematical model and the received data; and

allocating advertising space among available online advertisements based on the predicted future state of the aspect of the online advertisement.

47. The method of claim 46, wherein the aspect is a volume of impressions, clicks, or conversions within a given time period.

48. The method of claim 46, wherein the aspect is a performance rate of at least one of impression ratios, click ratios, or conversion ratios.

49. The method of claim 46, wherein advertising space is allocated by adjusting a control signal value associated with the aspect of the online advertisement.

50. A computer-implemented method for analyzing online advertisement-related data, the method comprising:

modeling an aspect of an online advertisement with a mathematical model, wherein the mathematical model comprises a control-signal-related component, and a control-signal-independent, time-based periodic component;

receiving data related to the aspect of the online advertisement;

predicting a future level of the aspect of the online advertisement, based on the mathematical model and the received data; and

choosing a control signal value based on a desired level for the aspect and the predicted future level of the aspect.

51. The method of claim 50, wherein the aspect is a volume of at least one of impressions, clicks, or conversions within a given time period.

52. The method of claim 50, wherein the aspect is a performance rate of impression ratios, click ratios, or conversion ratios.

53. The method of claim 50, wherein the control signal value affects an allocation of the online advertisement within an advertising space.