Qualification of website data and analysis using anomalies relative to historic patterns
View Patent ↗Tools and techniques are provided to assist detection of invalid clicks in website activity data. A system calculates or otherwise obtains a historic click quality characterization based on historic web analytics data. The system then identifies a click quality anomaly in the website activity data, namely, a departure from the historic click quality characterization. The identified anomalies may then be used to help guide searches for invalid clicks.
1. A method to assist detection of invalid clicks in network content site activity, the method comprising:
obtaining first and second historic click quality characterizations for a network content site for respective first and second time periods, wherein each of the first and second historic click quality characterizations is based on a comparison of clickthroughs and clickpasts and is obtained for each of the first and second time periods by:
receiving a clickpast threshold value that is determined based on at least one of:
(i) an average page depth of visits during the respective time period,
(ii) an average page length of visits, and
(iii) an average distribution of visits among different landing pages in a campaign,
determining a respective number of clickpasts for at least one content page for the respective time period, wherein each clickpast comprises a respective network content site visit involving clicking a respective number of pages past the at least one content page that is greater than the clickpast threshold value, wherein the respective number of clickpasts comprises at least one of (i) a respective number of unique pages clicked past the at least one content page and (ii) a respective number of successively deeper pages in the network content site clicked past the at least one content page, and
comparing the respective number of clickpasts with a respective number of clickthroughs for the at least one content page during the respective time period to obtain the historic click quality characterization; and
identifying a click quality anomaly based on a difference between the first and second historic click quality characterizations exceeding at least one of (i) a p-value indicative of a reliability of at least one of the first and second historic click quality characterizations and (ii) a standard deviation for at least one of the first and second historic click quality characterization.
2. The method of claim 1 , further comprising determining the clickpast threshold value based on at least two of the average page depth of visits, the average page length of visits, and the average distribution of visits among different landing pages.
3. The method of claim 1 , wherein each historic click quality characterization comprises a ratio of the number of clickpasts to the number of clickthroughs.
4. The method of claim 1 , wherein each historic click quality characterization is based at least in part on a measure of instances in which a network content site visitor clicks through to at least one campaign landing page at least two subsequent times in a row without also clicking at least one page past the at least one campaign landing page.
5. The method of claim 1 , wherein the average page depth of visits comprises an average number of pages clicked beyond the at least one content page of the network content site and the average page length of visits comprises an average number of pages visited in the network content site before exiting the network content site.
6. The method of claim 1 , wherein the at least one content page comprises at least one landing page of the network content site that is accessible via the campaign, wherein the campaign comprises electronic content other than a page of the network content site providing a link to the at least one landing page.
7. The method of claim 6 , wherein the campaign comprises at least one of a banner, an advertisement, a paid search engine result, and an e-mail communication.
8. A system comprising:
a processor; and
a non-transitory computer-readable medium in communication with the processor,
wherein the processor is configured for executing code stored in the non-transitory computer-readable medium to perform operations comprising:
obtaining first and second historic click quality characterizations for a network content site for respective first and second time periods, wherein each of the first and second historic click quality characterizations is based on a comparison of clickthroughs and clickpasts and is obtained for each of the first and second time periods by:
receiving a clickpast threshold value that is determined based on at least one of:
(i) an average page depth of visits comprising an average number of pages clicked beyond at least one content page of the network content site for visits during the respective time period,
(ii) an average page length of visits comprising an average number of pages visited in the network content site before exiting the network content site, and
(iii) an average distribution of visits among different landing pages in a campaign,
determining a respective number of clickpasts for the at least one content page for the respective time period, wherein each clickpast comprises a respective network content site visit involving clicking a respective number of pages past the at least one content page that is greater than the clickpast threshold value, wherein the respective number of clickpasts comprises at least one of (i) a respective number of unique pages clicked past the at least one content page and (ii) a respective number of successively deeper pages in the network content site clicked past the at least one content page, and
comparing the respective number of clickpasts with a respective number of clickthroughs for the at least one content page during the respective time period to obtain the historic click quality characterization; and
identifying a click quality anomaly based on a difference between the first and second historic click quality characterizations exceeding at least one of (i) a p-value indicative of a reliability of at least one of the first and second historic click quality characterizations and (ii) a standard deviation for at least one of the first and second historic click quality characterizations.
9. The system of claim 8 , wherein the operations further comprise determining the clickpast threshold value based on at least two of the average page depth of visits, the average page length of visits, and the average distribution of visits among different landing pages.
10. The system of claim 8 , wherein the at least one content page comprises at least one landing page of the network content site that is accessible via the campaign, wherein the campaign comprises electronic content other than a page of the network content site providing a link to the at least one landing page.
11. A non-transitory computer-readable storage medium storing instructions comprising:
instructions for obtaining first and second historic click quality characterizations for a network content site for respective first and second time periods, wherein each of the first and second historic click quality characterizations is based on a comparison of clickthroughs and clickpasts and is obtained for each of the first and second time periods by:
receiving a clickpast threshold value that is determined based on at least one of:
(i) an average page depth of visits comprising an average number of pages clicked beyond at least one content page of the network content site for visits during the respective time period,
(ii) an average page length of visits comprising an average number of pages visited in the network content site before exiting the network content site, and
(iii) an average distribution of visits among different landing pages in a campaign,
determining a respective number of clickpasts for the at least one content page for the respective time period, wherein each clickpast comprises a respective network content site visit involving clicking a respective number of pages past the at least one content page that is greater than the clickpast threshold value, wherein the respective number of clickpasts comprises at least one of:
a respective number of unique pages clicked past the at least one content page; and
a respective number of successively deeper pages in the network content site clicked past the at least one content page, and
comparing the respective number of clickpasts with a respective number of clickthroughs for the at least one content page during the respective time period to obtain the historic click quality characterization; and
instructions for identifying a click quality anomaly based on a difference between the first and second historic click quality characterizations exceeding at least one of (i) a p-value indicative of a reliability of at least one of the first and second historic click quality characterizations and (ii) a standard deviation for at least one of the first and second historic click quality characterizations.
12. The non-transitory computer-readable medium of claim 11 , further comprising program code for determining the clickpast threshold value based on at least two of the average page depth of visits, the average page length of visits, and the average distribution of visits among different landing pages.
13. The non-transitory computer-readable medium of claim 11 , wherein the at least one content page comprises at least one landing page of the network content site that is accessible via the campaign, wherein the campaign comprises electronic content other than a page of the network content site providing a link to the at least one landing page.