Generating actionable suggestions for improving user engagement with online advertisements

1. A computer system comprising:

a circuit to receive data defining an advertisement from an advertiser;

a feature extraction circuit configured to:

receive, based upon the data defining the advertisement, data defining an ad landing page associated with the advertisement;

identify a set of features of the ad landing page; and

produce a feature representation of the ad landing page associated with the advertisement and a set of ad categories based upon the set of features, wherein producing the feature representation comprises determining one or more values for one or more features including at least one of a first value indicative of a number of images on the ad landing page or a second value indicative of a number of links on the ad landing page;

a second circuit operative to receive historical dwell time data from one or more computers;

a dwell time prediction circuit operative to predict a distribution of dwell times for the advertisement using the set of ad categories and the historical dwell time data;

an ad quality rating generation circuit operative to compute a quality rating for the advertisement based upon (i) the feature representation, of the ad landing page associated with the advertisement, comprising at least one of the first value indicative of the number of images on the ad landing page or the second value indicative of the number of links on the ad landing page and (ii) the distribution of dwell times; and

a suggestion circuit operative to:

determine suggestions for improving the advertisement using the quality rating for the advertisement and a target range; and

provide the suggestions to the advertiser as feedback on ad performance to improve future performance of the advertisement.

2. The computer system of claim 1 wherein the feature representation comprises the first value indicative of the number of images on the ad landing page, wherein computing the quality rating is based upon the first value, of the feature representation, indicative of the number of images on the ad landing page.

3. The computer system of claim 1 wherein the feature representation comprises the second value indicative of the number of links on the ad landing page, wherein computing the quality rating is based upon the second value, of the feature representation, indicative of the number of links on the ad landing page.

4. The computer system of claim 1 wherein the ad quality rating generation circuit operates to identify a topical category of the ad landing page and operates to identify similar advertisements as advertisements having a same topical category.

5. The computer system of claim 4 comprising a communication circuit operative for data communication with a content analysis web service, wherein the computer system is operative to communicate the ad landing page to the content analysis web service and to receive the topical category of the ad landing page in return from the content analysis web service.

6. The computer system of claim 1 wherein providing the suggestions comprises providing, for display, a first actionable guideline and a second actionable guideline, wherein the first actionable guideline comprises first text indicative of a first suggestion and a first graphical representation of a first level of a first feature of the advertisement, wherein the second actionable guideline comprises second text indicative of a second suggestion and a second graphical representation of a second level of a second feature of the advertisement, wherein the first suggestion is associated with the first feature and the second suggestion is associated with the second feature.

7. The computer system of claim 1 wherein the feature extraction circuit operates to group features of the set of features of the ad landing page into two or more feature domains, and wherein the suggestion circuit operates to provide to the advertiser a set of suggestions, each suggestion of the set of suggestions corresponding to one feature domain of the two or more feature domains.

8. The computer system of claim 1 wherein providing the suggestions comprises providing, for display, a first actionable guideline and a second actionable guideline, wherein the first actionable guideline comprises a first graphical representation of a first level of a first feature of the advertisement, wherein the second actionable guideline comprises a second graphical representation of a second level of a second feature of the advertisement.

9. The computer system of claim 8 wherein the first graphical representation comprises a first color associated with a first level of importance and the second graphical representation comprises a second color associated with a second level of importance.

10. A method comprising:

receiving from an advertiser an advertisement to be displayed on an advertising system;

receiving historical dwell time data from one or more computers;

predicting a dwell time for the advertisement based upon the historical dwell time data;

extracting features associated with the advertisement, wherein the features comprise a feature representation of an ad landing page associated with the advertisement, wherein the feature representation comprises one or more values for one or more features including at least one of a first value indicative of a number of images on the ad landing page or a second value indicative of a number of links on the ad landing page;

determining a quality rating for the advertisement based upon (i) the dwell time for the advertisement and (ii) the feature representation of the ad landing page associated with the advertisement;

determining suggestions for improving the advertisement using the quality rating for the advertisement; and

providing the suggestions to the advertiser.

11. The method of claim 10 wherein the feature representation comprises the first value indicative of the number of images on the ad landing page.

12. The method of claim 10 wherein the feature representation comprises the second value indicative of the number of links on the ad landing page, wherein determining the quality rating is based upon the second value, of the feature representation, indicative of the number of links on the ad landing page.

13. The method of claim 10 wherein providing suggestions to the advertiser for improvements comprises providing to the advertiser a list of actionable guidelines that can improve expected dwell time of the advertisement.