Audience duplication for parent-child resource pairs
A combined reach can be determined for multiple resources. Usage data, including audience duplication data, is used to generate audience duplication measurements for pairs of the resources. A child-resource pair is identified, and the audience duplication measurement for that resource pair is modified. The modified measurement is used to determine a combined reach for the multiple resources. A report is also generated based on the combined reach.
1. A computer-implemented method comprising:
accessing, by one or more processors, usage data that indicates accesses to a plurality of different resources by one or more members of an audience with one or more client systems and includes information about audience duplication across one or more of the plurality of different resources;
generating, based on the usage data and by the one or more processors, audience duplication measurements for pairs of resources in the plurality of different resources;
determining, by the one or more processors, that at least one of the pairs of resources is a parent-child resource pair, wherein a parent-child resource pair includes two resources with a parent-child relationship such that a first one of the two resources is a parent resource and a second one of the two resources is a child resource, the parent-child resource pair having an associated audience duplication measurement generated based on the usage data;
in response to determining that at least one of the pairs of resources is a parent-child resource pair, modifying, by the one or more processors, the audience duplication measurement associated with the parent-child resource pair to generate a modified audience duplication measurement;
determining, by the one or more processors, a combined reach representing a number of accesses of the plurality of different resources by unique members of the audience based at least on the usage data and the modified audience duplication measurement; and
generating, by the one or more processors, a report based on the combined reach.
2. The computer-implemented method of claim 1 , wherein the information about audience duplication included in the accessed usage data is correlation data calculated between each pair of the plurality of different resources, and wherein the audience duplication measurement associated with the parent-child resource pair is a pairwise correlation coefficient for the parent-child resource pair.
3. The computer-implemented method of claim 2 , wherein modifying the pairwise correlation coefficient for the parent-child resource pair comprises multiplying the pairwise correlation coefficient by a factor based on a ratio between usage of the child resource and usage of the parent resource.
4. The computer-implemented method of claim 3 , wherein the factor is proportional to (1−p c /p p ), wherein p c is the usage of the child resource and p p is the usage of the parent resource.
5. The computer-implemented method of claim 3 , further comprising calculating an adjusted combined reach based on a product of a measured unique audience for each of the plurality of different resources multiplied by an adjustment factor, the adjustment factor being calculated using the modified pairwise correlation coefficients.
6. The computer-implemented method of claim 1 , wherein determining that the parent-child relationship exists comprises identifying that the audience duplication of usage of the parent-child resource pair is equal to usage of the child resource.
7. The computer-implemented method of claim 1 , wherein the child resource comprises one or more websites, and wherein the parent resource comprises all of the websites of the child resource and one or more other websites.
8. The computer-implemented method of claim 1 , wherein the plurality of different resources are hosted on servers, and wherein the one or more members of the audience access the plurality of different resources using client systems that send requests for the plurality of different resources to the servers across a network.
9. The computer-implemented method of claim 1 , wherein the combined reach is also determined based on one or more of the generated audience duplication measurements.
10. The computer-implemented method of claim 1 , wherein the modification of the audience duplication measurement takes into account an observed negative relationship between a parent-child resource pair and combined reach such that the combined reach determined based on the modified audience duplication measurement is less than a combined reach determined using the same data but the unmodified audience duplication measurement.
11. A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
accessing usage data that indicates accesses to a plurality of different resources by one or more members of an audience with one or more client systems and includes information about audience duplication across one or more of the plurality of different resources;
generating, based on the usage data, audience duplication measurements for pairs of resources in the plurality of different resources;
determining that at least one of the pairs of resources is a parent-child resource pair, wherein a parent-child resource pair includes two resources with a parent-child relationship such that a first one of the two resources is a parent resource and a second one of the two resources is a child resource, the parent-child resource pair having an associated audience duplication measurement generated based on the usage data;
in response to determining that at least one of the pairs of resources is a parent-child resource pair, modifying the audience duplication measurement associated with the parent-child resource pair to generate a modified audience duplication measurement;
determining a combined reach representing a number of accesses of the plurality of different resources by unique members of the audience based at least on the usage data and the modified audience duplication measurement; and
generating a report based on the combined reach.
12. The system of claim 11 , wherein the information about audience duplication included in the accessed usage data is correlation data calculated between each pair of the plurality of different resources, and wherein the audience duplication measurement associated with the parent-child resource pair is a pairwise correlation coefficient for the parent-child resource pair.
13. The system of claim 12 , wherein modifying the pairwise correlation coefficient for the parent-child resource pair comprises multiplying the pairwise correlation coefficient by a factor based on a ratio between usage of the child resource and usage of the parent resource.
14. The system of claim 13 , wherein the factor is proportional to (1−p c /p p ), wherein p c is the usage of the child resource and p p is the usage of the parent resource.
15. The system of claim 13 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising calculating an adjusted combined reach based on a product of a measured unique audience for each of the plurality of different resources multiplied by an adjustment factor, the adjustment factor being calculated using the modified pairwise correlation coefficients.
16. The system of claim 11 , wherein determining that the parent-child relationship exists comprises identifying that the audience duplication of usage of the parent-child resource pair is equal to usage of the child resource.
17. The system of claim 11 , wherein the child resource comprises one or more websites, and wherein the parent resource comprises all of the websites of the child resource and one or more other websites.
18. The system of claim 11 , wherein the plurality of different resources are hosted on servers, and wherein the one or more members of the audience access the plurality of different resources using client systems that send requests for the plurality of different resources to the servers across a network.
19. The system of claim 11 , wherein the combined reach is also determined based on one or more of the generated audience duplication measurements.
20. The system of claim 11 , wherein the modification of the audience duplication measurement takes into account an observed negative relationship between a parent-child resource pair and combined reach such that the combined reach determined based on the modified audience duplication measurement is less than a combined reach determined using the same data but the unmodified audience duplication measurement.
21. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
accessing usage data that indicates accesses to a plurality of different resources by one or more members of an audience with one or more client systems and includes information about audience duplication across one or more of the plurality of different resources;
generating, based on the usage data, audience duplication measurements for pairs of resources in the plurality of different resources;
determining that at least one of the pairs of resources is a parent-child resource pair, wherein a parent-child resource pair includes two resources with a parent-child relationship such that a first one of the two resources is a parent resource and a second one of the two resources is a child resource, the parent-child resource pair having an associated audience duplication measurement generated based on the usage data;
in response to determining that at least one of the pairs of resources is a parent-child resource pair, modifying the audience duplication measurement associated with the parent-child resource pair to generate a modified audience duplication measurement;
determining a combined reach representing a number of accesses of the plurality of different resources by unique members of the audience based at least on the usage data and the modified audience duplication measurement; and
generating a report based on the combined reach.