Estimating inventory, user behavior, and/or cost and presentation attributes for an advertisement for use with an advertising system

• US 8,645,192 B1
• Filed: 07/16/2012
• Issued: 02/04/2014
• Est. Priority Date: 01/10/2003
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Sandberg, Magnus, Veach, Eric, Qian, Zhe
• Primary Examiner(s): CHAMPAGNE, DONALD
• Application Number:
  US13/550,263
• Time in Patent Office: 568 Days
• Field of Search: 705/7.35, 705/14.54, 705/14.71
• US Class Current: 705/7.35
• CPC Class Codes: G06Q 30/0206 Price or cost determination based on market factors
• CPC Class Codes: G06Q 30/0243 Comparative campaigns

Abstract

An advertiser may be provided with fast and accurate estimates of the future performance of at least one of their advertisements. These estimates may be broken down by keyword, thereby providing, up-front, useful information to the advertiser about how different keywords will affect the performance and cost of their ad campaign. Estimates and/or suggestions for a set of keywords, given additional campaign parameters, may be generated by using historical search traffic (or some other ad serving trigger) data to predict future inventory and to find useful keyword (or some other serving constraint) variations. Current data on competing ads may be used to simulate the competitions (e.g., real-time auctions) that the given ads will participate in.

Claims

1. A computer-implemented method comprising:

identifying, using a computer system including one or more computers, a candidate ad;

determining, using the computer system, a plurality of served search queries received from a plurality of users that could trigger the serving of the candidate ad;

determining, using the computer system, for at least some of the plurality of served search queries, at least one other ad that could be triggered by the served search queries;

simulating, using the computer system and for each of the plurality of served search queries, an auction between the at least one other ad and the candidate ad to generate simulation data; and

determining, using the computer system, cost data based on the simulation data that reflects an expected cost to be associated with the candidate ad that will result in a predetermined level of performance for the candidate.

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

estimating, using the computer system, an inventory of ad impressions available to the candidate ad over a predetermined time period, wherein the auctions simulated are for the estimated inventory.

3. The computer-implemented method of claim 2 wherein the act of estimating the inventory of ad impressions available to the candidate ad over the predetermined time period includes

obtaining a set of potentially matching ad serving trigger samples,

classifying the ad serving triggers of the obtained set,

tallying a best matching serving constraint for each of the ad serving triggers of the obtained set,

combining equivalent ad serving triggers of the obtained set to generate an updated set,

extracting ad serving trigger, frequency pairs from the updated set,

sorting the pairs, and

determining an inventory estimate for each serving constraint associated with the candidate ad.

4. The computer-implemented method of claim 3 wherein the ad serving triggers are a second plurality of served search queries received from a plurality of users.

5. The computer-implemented method of claim 3 wherein the serving constraints are keywords used to target the serving of the candidate ad.

6. The computer-implemented method of claim 3 wherein the ad serve trigger samples are samples of a second plurality of served search queries received from a plurality of users that could have triggered serving of the candidate ad.

7. The computer-implemented method of claim 2 wherein estimating the inventory of ad impressions available to the candidate ad over the predetermined time period includes

obtaining a set of potentially matching ad serving trigger samples,

classifying the ad serving triggers of the obtained set,

tallying a best matching serving constraint for each of the ad serving triggers of the obtained set, and

determining an inventory estimate for each serving constraint associated with the candidate ad.

8. The computer-implemented method of claim 7 wherein the ad serving triggers are a second plurality of served search queries received from a plurality of users.

9. The computer-implemented method of claim 7 wherein the serving constraints are keywords used to target the serving of the candidate ad.

10. The computer-implemented method of claim 7 wherein the ad serving trigger samples are samples of a second plurality of served search queries received from a plurality of users that could have triggered serving of the candidate ad.

11. The computer-implemented method of claim 1 wherein the expected cost is a cost per impression.

12. The computer-implemented method of claim 1 wherein the expected cost is a cost per selection.

13. The computer-implemented method of claim 1 further comprising determining based on the simulation data presentation attribute information associated with each auction wherein the presentation attribute is a position relative to other served ads in a simulated auction, the method further comprising providing the presentation attribute information along with the expected cost.

14. The computer-implemented method of claim 1 wherein simulating auctions includes, for each serving constraint of the candidate ad,

accepting the plurality of served search queries received from a plurality of users matching the serving constraint;

for each of the accepted plurality of served search queries, using the accepted plurality of served search queries to simulate an auction between the candidate ad and competing ads to obtain at least one of a cost attribute sample and a presentation attribute sample, and

determining at least one of a composite cost using the cost samples and presentation attribute samples.

15. The computer-implemented method of claim 14 wherein the serving constraints of the candidate ad are keywords used to trigger the serving of the candidate ad.

16. The computer-implemented method of claim 14 wherein redundant ones of the plurality of served search queries are combined and assigned a frequency of occurrence, wherein the plurality of served search queries are ordered based on the frequency of occurrence.

17. Apparatus comprising;

one or more processors;

at least one input device; and

one or more storage devices storing processor-executable instructions which, when executed by one or more processors, perform a method of;

identifying a candidate ad;

determining a plurality of served search queries received from a plurality of users that would trigger the serving of the candidate ad;

determining for at least some of the plurality of served search queries, at least one other ad that could be triggered by the served search queries;

simulating, for each of the plurality of served search queries, an auction between the at least one other ad and the candidate ad to generate simulation data; and

determining an expected cost to be associated with the candidate ad to meet a predefined performance criteria.

18. The apparatus of claim 17 wherein the processor-executable instructions, when executed by one or more processors, further perform a method of;

estimating an inventory of ad impressions available to the candidate ad over a predetermined time period, wherein the auctions simulated are for the estimated inventory.

19. A computer-implemented method comprising:

identifying, using a computer system including one or more computers, a candidate ad;

determining, using the computer system and using historical ad serving triggers that could have triggered the serving of the candidate ad, a predicted future inventory of ad impressions for which the candidate ad would be eligible to compete;

simulating, using the computer system, for at least one ad impression of the predicted future inventory of ad impressions, an auction between the candidate ad and a set of competing ads for serving on the predicted future inventory of ad impressions; and

determining, using the computer system, presentation attribute data based on simulation data derived from the simulating that reflects an expected presentation attribute to be associated with the candidate ad that will result based on a predefined cost per result bid for the candidate in the auctions.

20. The computer-implemented method of claim 19 wherein the competing ads are presently existing ads.

21. The computer-implemented method of claim 19 further comprising:

tracking, using the computer system, for each of the simulated auctions between the candidate ad and set of competing ads, at least one of (A) a cost of serving the candidate ad, or (B) a presentation attribute of the candidate ad.

22. The computer-implemented method of claim 21 further comprising:

determining, using the computer system, for each of the simulated auctions between the candidate ad and set of competing ads, a composite cost of serving the candidate ad.

23. The computer-implemented method of claim 22 wherein the composite cost is a cost per impression.

24. The computer-implemented method of claim 22 wherein the composite cost is a cost per selection.

25. The computer-implemented method of claim 21 further comprising:

determining, using the computer system, for each of the simulated auctions between the candidate ad and set of competing ads, a composite presentation attribute of the candidate ad.

26. The computer-implemented method of claim 22 wherein the presentation attribute is a position relative to other served ads.

27. The computer-implemented method of claim 19 wherein the ad serving triggers are a plurality of served search queries received from a plurality of users.

28. The computer-implemented method of claim 27 wherein the historical ad serving triggers that could have triggered the serving of the candidate ad are taken from a sample of the plurality of served search queries received from the plurality of users.

29. Apparatus comprising;

one or more processors;

at least one input device; and

one ore more storage devices storming processor-executable instructions which, when executed by one or more processors, perform a method of;

identifying a candidate ad;

determining, using historical ad serving triggers that could have triggered the serving of the candidate ad, a predicted future inventory of ad impressions for which the candidate ad would be eligible to compete;

simulating, for at least one ad impression of the predicted future inventory of ad impressions, an auction between the candidate ad and a set of competing ads for serving on the predicted future inventory of ad impressions; and

determining, using the computer system, presentation attribute data based on simulation data derived from the simulating that reflects an expected presentation attribute to be associated with the candidate ad that will result based on a predefined cost per result bid for the candidate in the auctions.

30. The apparatus of claim 29 wherein the processor-executable instructions, when executed by one or more processors, further perform a method of:

tracking, for each of the simulated auctions between the candidate ad and set of competing ads, at least one of (A) a cost of serving the candidate ad, or (B) a presentation attribute of the candidate ad.

31. The apparatus of claim 30 wherein the processor-executable instructions, when executed by one or more processors, further perform a method of:

determining a composite cost of serving the candidate ad.

32. The apparatus of claim 30 wherein the processor-executable instructions, when executed by one or more processors, further perform a method of:

determining a composite presentation attribute of the candidate ad.

33. The apparatus of claim 29 wherein the ad serving triggers are a plurality of served search queries received from a plurality of users.

34. The apparatus of claim 33 wherein the historical ad serving triggers that could have triggered the serving of the candidate ad are taken from a sample of the plurality of served search queries received from the plurality of users.