Considering user-relevant criteria when serving advertisements

• US 8,533,044 B2
• Filed: 04/14/2010
• Issued: 09/10/2013
• Est. Priority Date: 04/14/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Oren, Ishai, Aharoni, Asaf, Portugaly, Elon, Perri, Tal, Zilberstein, Benny
• Primary Examiner(s): VANDERHORST, MARIA VICTORIA
• Application Number:
  US12/760,299
• Publication Number:
  US 20110258035A1
• Time in Patent Office: 1,245 Days
• Field of Search: 705/14.42, 705/14.43, 705/14.41
• US Class Current: 705/14.43
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0244 Optimization

Abstract

The present invention is directed to selecting an advertisement to be presented in an ad space of a webpage. A criterion (e.g., keyword, image, audio element, etc.) is selected to evoke the advertisement. In an embodiment, the criterion is selected based on a relevance of the criteria to a user, and not necessarily based solely on a relevance to the webpage. Whether a criterion should be used to evoke the advertisement might be based on various factors, such as the likelihood that the user will select a criterion-evoked advertisement, regardless of whether the criterion appears in among content of the webpage, and an expected gain of presenting the criterion-evoked advertisement.

Claims

1. A computer-readable hardware memory device storing computer-executable instructions that, when executed, cause a computing device to perform a method of selecting an advertisement to be displayed on a webpage, the method comprising:

receiving the webpage that is to be served to a recipient device and that includes a fillable advertisement space, wherein the webpage includes a first keyword among content of the webpage;

retrieving a user-behavior click-through-rate (uCTR) of a second keyword, which was available to evoke a prior advertisement previously served to a user of the recipient device and is not included among the content of the webpage,

wherein the uCTR quantifies a relevance of the second keyword to the user, and

wherein the uCTR is calculated using a formula that mathematically combines a measured click-through-rate (mCTR) with a set of learned coefficients, the mCTR quantifying a contextual relevance of the second keyword to the user in a context of when the prior webpage was served to the user; and

based at least in part on the uCTR, using the second keyword to evoke the advertisement instead of the first keyword, wherein the second keyword is not included among the content of the webpage and the first keyword is include among the content of the webpage.

2. Computer-readable hardware memory device of claim 1, wherein the uCTR suggests whether the advertisement served with the webpage will be selected by the user when the advertisement is evoked by the second keyword.

3. Computer-readable hardware memory device of claim 1,

wherein the prior advertisement was served in a prior advertisement space of a prior webpage;

wherein the uCTR is calculated by applying a link function the mCTR, such that the uCTR quantifies a learned relevance of the second keyword to the user, and

wherein the mCTR quantifies a contextual relevance of the second keyword to the user in a context of when the prior webpage was served to the user.

4. Computer-readable hardware memory device of claim 3,

wherein the uCTR is determined using a formula represented by:

uCTR=(az+b)/(cz+d);

wherein z represents the mCTR of a pairing of the prior advertisement space and the second keyword; and

wherein a, b, c, and d represent learned coefficients.

5. Computer-readable hardware memory device of claim 1, wherein the second keyword is available to evoke the prior advertisement when the second keyword is included among content of a prior webpage that was served to the user together with the prior advertisement.

6. Computer-readable hardware memory device of claim 1 further comprising, conducting an auction between the second keyword and the first keyword,

wherein a first auction rank of the second keyword is equal to a product of a first bid and the uCTR, and

wherein the first rank is more favorable than a second rank of the first keyword, such that, based at least in part on the uCTR, the second keyword is used to evoke the advertisement.

7. A method, which is facilitated by a processor and computer-readable media, of selecting an advertisement to be displayed on a webpage, the method comprising:

serving to a recipient device an initial webpage having an initial advertisement and a webpage keyword, which is included among text of the initial webpage, wherein the initial advertisement is served in an initial fillable advertisement space and is not evoked by the webpage keyword;

calculating by the processor a measured click-through-rate (mCTR) that quantifies a relevance of the webpage keyword to a recipient-device user in a context of when the initial webpage is served with the initial advertisement;

using the processor to calculate a user-behavior click-through-rate (uCTR) by applying a formula that mathematically combines the mCTR with a set of learned coefficients, the uCTR suggesting, whether a subsequent advertisement served with a subsequent webpage will be selected by the recipient-device user when the subsequent advertisement is evoked by the webpage keyword; and

conducting by the processor an auction between the webpage keyword and a current-webpage keyword, which is included among the text of the subsequent webpage,

wherein a result of the auction is at least in part determined using the uCTR of the webpage keyword, and

wherein, based on the result, the webpage keyword is used, instead of the current-webpage keyword, to evoke the subsequent advertisement, which is served to the recipient-device user.

8. The method of claim 7,

wherein, when the initial advertisement is evoked by the webpage keyword and the initial advertisement is not selected by the recipient-device user, the mCTR is represented by a first ratio:

(SEL/IMP),

SEL being a number of times an advertisement evoked by the webpage keyword is selected by any user of any recipient device when the advertisement is served in the initial fillable advertisement space, and

IMP being a number of times the advertisement is served to any recipient device in the initial fillable advertisement space; and

wherein, when the initial advertisement is not evoked by the webpage keyword, but the webpage keyword was a source usable to evoke an alternative advertisement, the mCTR is calculated using a first formula represented by;

mCTR=(SEL/IMP)*(IMP/SOU)^α

,

SOU being equal to the number of times the webpage keyword was a source to evoke the alternative advertisement, regardless of whether the alternative advertisement was served, and

α

being a parameter that is learned.

9. The method of claim 8, wherein, when the initial advertisement is not evoked by the webpage keyword and the webpage keyword was not a source to evoke the alternative advertisement, the mCTR is calculated using a second formula represented by:

mCTR=((SEL/IMP)*(IMP/SOU)^α

)*(SOU/TOT)^β

,

wherein TOT is equal to a number of times the initial fillable advertisement space was served to any user of any recipient device, and

wherein β

is a parameter that is learned.

10. The method of claim 7 further comprising, ranking the webpage keyword as a user keyword among other user keywords, wherein a keyword score of the user keyword is determined using the uCTR and wherein the keyword score is used to rank the user keyword.

11. The method of claim 10, wherein the keyword score comprises a product of the uCTR and a cost-per-click value of serving a given advertisement evoked by the user keyword.

12. The method of claim 11, wherein the keyword score comprises the product minus an expected opportunity cost of not serving a contextual advertisement on a given webpage.

13. The method of claim 7,

wherein the result is determined by assigning a respective auction rank to each of the webpage keyword and the current-webpage keyword,

wherein a first auction rank of the webpage keyword is equal to a product of a first bid and the uCTR, and

wherein the first auction rank is more favorable than a second auction rank of the current-webpage keyword, such that, based at least in part on the uCTR, the webpage keyword is used to evoke the subsequent advertisement.

14. The method of claim 13,

wherein the second auction rank is equal to a product of a second bid, a predicted click-through-rate (pCTR), and a measure of relevance of the current-webpage keyword to the subsequent webpage, and

wherein the pCTR quantifies a general relevance between the current-webpage keyword and a potential advertisement evoked by the current-webpage keyword, the general relevance not taking into account a relevance of the current-webpage keyword to the recipient-device user.

15. A system, which is employed using a processor and computer-readable media, that selects an advertisement to be displayed on a webpage, the system comprising:

a user-criteria database storing a user-associated criterion together with a user-behavior click-through-rate (uCTR),

wherein the user-associated criterion was available to evoke an initial advertisement that was previously served together with an initial webpage to a user, and

wherein the uCTR suggests whether a subsequent advertisement served with a subsequent webpage will be selected by the user when the subsequent advertisement is evoked by the user-associated criterion;

a criteria-stat generator that calculates the uCTR using a formula that mathematically combines a measured click-through-rate (mCTR) with a set of learned coefficients, the mCTR quantifying a relevance of the user-associated criterion to a recipient-device user in a context of when the initial webpage is served with the initial advertisement; and

a criteria evaluator that conducts an auction between the user-associated criterion and a current-webpage criteria, which is included in among content of the subsequent webpage.

16. The system of claim 15, wherein a value of the mCTR is dependent on whether the user-associated criterion was used to evoke the initial advertisement.

17. The system of claim 16, wherein, when the user-associated criterion was used to evoke the initial advertisement, the value of the mCTR is dependent on whether the initial advertisement was selected by the user.