Conversion crediting
Methods, systems, and apparatus, including computer program products, for processing events related to presented content. In one aspect, a method includes determining a time window count of a number of advertising events associated with an advertisement during at least one time window before a conversion event; and determining a credit that represents a strength of an association between the advertisement and the conversion event, wherein determining a credit includes selecting a weighting model for the at least one time window count.
1. A computer-implemented method, comprising:
identifying, by one or more computers, user interactions with a particular advertisement;
identifying, by the one or more computers, multiple conversion events associated with the particular advertisement, each identified conversion event corresponding to one of the user interactions, wherein the identified conversion event is a predefined interaction with a resource;
identifying, by the one or more computers and for each of the identified multiple conversion events, a conversion time that specifies an amount of time elapsed between the identified conversion event and the user interaction to which the identified conversion event corresponds;
generating, by the one or more computers, a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified multiple conversion events, and each conversion time, wherein the time-to-conversion profile indicates a likelihood of a conversion event at multiple different amounts of time elapsed after a user interaction associated with the particular advertisement;
determining, by the one or more computers and based on the time-to-conversion profile, multiple different conversion time windows having sizes that are based on different slopes of the time-to-conversion profile, wherein a first time window determined for a first portion of the time-to-conversion profile having a first slope is smaller than a second time window for a second portion of the time-to-conversion profile having a second slope that is less than the first slope; and
outputting, by the one or more computers, data that graphically present the time-to-conversion profile to a provider of the particular advertisement in a reporting interface.
2. The method of claim 1 , wherein generating, by the one or more computers, a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified conversion events, and each conversion time comprises:
for each of a plurality of time intervals, dividing a number of times an identified user interaction occurred before a corresponding identified conversion event in the time interval by an average number of user interaction occurrences in the time interval.
3. The method of claim 1 , further comprising:
determining, for each given time window among the multiple different conversion time windows, a corresponding time window count by determining a number of the identified user interactions occurring within the given time window and before the corresponding identified conversion event;
selecting a weighting factor for each given time window;
generating, for each given time window, a weighted time window count based on a function of each of the corresponding time window counts and the weighting factor for the given time window, the weighted time window count indicating a value of the identified user interactions that occurred within the given time window relative to the identified user interactions occurring in the other ones of the multiple different time windows; and
determining an advertisement credit for the particular advertisement based at least in part on the weighted time window counts for each given time window, the advertisement credit indicating a strength of an association between the particular advertisement and the identified conversion events.
4. The method of claim 3 , further comprising:
generating a normalized credit by multiplying the credit by a normalization factor, the normalization factor being such that a sum of normalized credits equals a total number of conversion events.
5. The method of claim 4 , further comprising:
determining a cost effectiveness of the advertisement by dividing a cost of the advertisement by the normalized credit for the advertisement.
6. A system, comprising:
a data processing apparatus; and
a data store storing instructions that, when executed by the data processing apparatus, cause the data processing apparatus to perform operations comprising:
identifying user interactions with a particular advertisement;
identifying multiple conversion events associated with the particular advertisement, each identified conversion event corresponding to one of the user interactions, wherein the identified conversion event is a predefined interaction with a resource;
identifying, for each of the identified multiple conversion events, a conversion time that specifies an amount of time elapsed between the identified conversion event and the user interaction to which the identified conversion event corresponds;
generating a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified conversion events, and each conversion time, wherein the time-to-conversion profile indicates a likelihood of a conversion event at multiple different amounts of time elapsed after a user interaction associated with the particular advertisement;
determining, based on the time-to-conversion profile, multiple different conversion time windows having sizes that are based on different slopes of the time-to-conversion profile, wherein a first time window determined for a first portion of the time-to-conversion profile having a first slope is smaller than a second time window for a second portion of the time-to-conversion profile having a second slope that is less than the first slope; and
outputting, by the one or more computers, data that graphically present the time-to-conversion profile to a provider of the particular advertisement in a reporting interface.
7. The system of claim 6 , wherein generating a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified conversion events, and each conversion time comprises:
for each of a plurality of time intervals, dividing a number of times an identified user interaction occurred before a corresponding identified conversion event in the time interval by an average number of user interaction occurrences in the time interval.
8. The system of claim 6 , wherein the operations further comprise:
determining, for each given time window among the multiple different time windows, a corresponding time window count by determining a number of the identified user interactions occurring within the given time window and before the corresponding identified conversion event;
selecting a weighting factor for each given time window;
generating, for each given time window, a weighted time window count based on a function of each of the corresponding time window counts and the weighting factor for the given time window, the weighted time window count indicating a value of the identified user interactions that occurred within the given time window relative to the identified user interactions occurring in the other ones of the multiple different time windows; and
determining an advertisement credit for the particular advertisement based at least in part on the weighted time window counts for each given time window, the advertisement credit indicating a strength of an association between the particular advertisement and the identified conversion events.
9. The system of claim 8 , wherein the operations further comprise:
generating a normalized credit by multiplying the credit by a normalization factor, the normalization factor being such that a sum of normalized credits equals a total number of conversion events.
10. The system of claim 9 , wherein the operations further comprise:
determining a cost effectiveness of the advertisement by dividing a cost of the advertisement by the normalized credit for the advertisement.
11. A non-transitory computer readable medium comprising instructions that, when executed by data processing apparatus, cause the data processing apparatus to perform operations comprising:
identifying user interactions with a particular advertisement;
identifying multiple conversion events associated with the particular advertisement, each identified multiple conversion event corresponding to one of the user interactions, wherein the identified conversion event is a predefined interaction with a resource;
identifying, for each of the identified conversion events, a conversion time that specifies an amount of time elapsed between the identified conversion event and the user interaction to which the identified conversion event corresponds;
generating a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified multiple conversion events, and each conversion time, wherein the time-to-conversion profile indicates a likelihood of a conversion event at multiple different amounts of time elapsed after a user interaction associated with the particular advertisement;
determining, based on the time-to-conversion profile, multiple different conversion time windows having sizes that are based on different slopes of the time-to-conversion profile, wherein a first time window determined for a first portion of the time-to-conversion profile having a first slope is smaller than a second time window for a second portion of the time-to-conversion profile having a second slope that is less than the first slope; and
outputting, by the one or more computers, data that graphically present the time-to-conversion profile to a provider of the particular advertisement in a reporting interface.
12. The non-transitory computer readable medium of claim 11 , wherein generating a time-to-conversion profile for the particular advertisement based on the identified user interactions, the identified conversion events, and each conversion time comprises:
for each of a plurality of time intervals, dividing a number of times an identified user interaction occurred before a corresponding identified conversion event in the time interval by an average number of user interaction occurrences in the time interval.
13. The non-transitory computer readable medium of claim 11 , wherein the operations further comprise:
determining, for each given time window among the multiple different conversion time windows, a corresponding time window count by determining a number of the identified user interactions occurring within the given time window and before the corresponding identified conversion event;
selecting a weighting factor for each given time window;
generating, for each given time window, a weighted time window count based on a function of each of the corresponding time window counts and the weighting factor for the given time window, the weighted time window count indicating a value of the identified user interactions that occurred within the time window relative to the identified user interactions occurring in the other ones of the multiple different time windows; and
determining an advertisement credit for the particular advertisement based at least in part on the weighted time window counts for each given time window, the advertisement credit indicating a strength of an association between the particular advertisement and the identified conversion events.
14. The non-transitory computer readable medium of claim 13 , wherein the operations further comprise:
generating a normalized credit by multiplying the credit by a normalization factor, the normalization factor being such that a sum of normalized credits equals a total number of conversion events.