Determining Device Counts

1. A system comprising: one or more processing devices; one or more storage devices storing instructions that, when executed by the one or more processing devices, cause the one or more processing devices to perform the following operations: access a set of activity data for a set of one or more resources on a network, the set of one or more resources having been accessed by a group of client systems and the set of activity data being determined based on beacon messages received from the group of client systems sent as a result of beacon instructions included with the first set of resources, wherein the set of activity data represents cookies received with the beacon messages and a total count of the beacon messages received from the group of client systems; determine persistent cookie data based on the set of activity data, the persistent cookie data representing multiple persistent cookies and, for each persistent cookie, a total count of beacon messages received from a client system associated with the persistent cookie, wherein each persistent cookie is associated with a different client system from the group of client systems and is received with the beacon messages that span a threshold amount of time during a reporting period; determine weighted count data based on the persistent cookie data, the weighted count data representing the multiple persistent cookies and, for each persistent cookie, an estimated count of beacon messages that would have been received with the persistent cookie from the client system associated with the persistent cookie if the persistent cookie persisted on the client system during the entire reporting period; determine a count per device value based on the weighted count data, the count per device value indicating an average number of beacon messages received per device during the reporting period; and determine a count of unique devices that have accessed the set of resources during the reporting period based on the count per device value and the total count of beacon messages received from the group of client systems.

2. The system of claim 1 wherein, to determine the persistent cookie data, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to: determine that a difference between a first time a cookie is received with a beacon message and a second time the cookie is received with a beacon message is greater than the threshold amount of time; and in response to determining that the difference is greater than the threshold, identify the cookie as one of the multiple persistent cookies.

3. The system of claim 2 wherein the reporting period is a month and the threshold amount of time is a threshold number of days.

4. The system of claim 1 wherein: the reporting period is divided into subunits and the threshold amount of time is a threshold fraction of subunits in the reporting period, and to determine the persistent cookie data, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to: determine that a fraction of the subunits in the reporting period in which a cookie is received with a beacon message exceeds the threshold fraction of subunits in the reporting period; and in response to determining that the fractions exceeds the threshold fraction of subunits in the reporting period, identify the cookie as the persistent cookie.

5. The system of claim 1 wherein, to determine the weighted count data, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to: for a given persistent cookie: determine a total time present in the reporting period, wherein the total time present is a time in the reporting period that the given persistent cookie was received with a beacon message; determine a total time not present in the reporting period, wherein the total time not present is a time in the reporting period that the given persistent cookie was not received with a beacon message; determine a total number of times the given persistent cookie was received during the reporting period with a beacon message; and determine an estimated count of beacon messages that would have been received with the given persistent cookie from a client device associated with the persistent cookie if the persistent cookie persisted on the client device during the entire reporting period, based on the total time present, the total time not present, and the total number of times the given persistent cookie was received during the reporting period with a beacon message.

6. The system of claim 1 wherein, to determine the weighted count data, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to: for a given persistent cookie: determine an active persistency value, wherein the active persistency value is a percentage of the reporting period in which a beacon message with the given persistent cookie was received; determine an active subunit value, wherein the active subunit value is a count of subunits of the reporting period during which a beacon message with the given persistent cookie was received; determine a total time not present in the reporting period, wherein the total time not present is a time in the reporting period that the given persistent cookie was not received with a beacon message; determine a total number of times the given persistent cookie was received during the reporting period with a beacon message; and determine an estimated count of beacon messages that would have been received with the given persistent cookie from a client device associated with the persistent cookie if the persistent cookie persisted on the client device during the entire reporting period, based on the activity persistence value, the active subunit value, the total time not present, and the total number of times the given persistent cookie was received during the reporting period with a beacon message.

7. The system of claim 1 wherein, to determine the count per device value, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to determine the count per device value based on the estimated count of beacon messages and a total number of persistent cookies in the multiple persistent cookies.

8. The system of claim 1 wherein the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to subdivide the persistent cookie data into one or more groupings, wherein each grouping corresponds to a category of client system; to determine the weighted count data the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to determine the weighted count data for each of the groupings; to determine the count per device value the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to determine the count per device value for each grouping based on the weighted count data for the grouping; and to determine the count of unique devices the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to determine the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received from the group of client systems.

9. The system of claim 8 wherein, to determine the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received from the group of client systems, the instructions include instructions that, when executed by the one or more processing devices, cause the one or more processing devices to: subdivide the total count of beacon messages received from the group of client systems into the groupings; and determine the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received for the grouping.

10. The system of claim 8 wherein the groupings are based on type of client system, operating system of the client system, or provider of connectivity of the client system.

11. A method comprising: accessing a set of activity data for a set of one or more resources on a network, the set of one or more resources having been accessed by a group of client systems and the set of activity data being determined based on beacon messages received from the group of client systems sent as a result of beacon instructions included with the first set of resources, wherein the set of activity data represents cookies received with the beacon messages and a total count of the beacon messages received from the group of client systems; determining persistent cookie data based on the set of activity data, the persistent cookie data representing multiple persistent cookies and, for each persistent cookie, a total count of beacon messages received from a client system associated with the persistent cookie, wherein each persistent cookie is associated with a different client system from the group of client systems and is received with the beacon messages that span a threshold amount of time during a reporting period; determining weighted count data based on the persistent cookie data, the weighted count data representing the multiple persistent cookies and, for each persistent cookie, an estimated count of beacon messages that would have been received with the persistent cookie from the client system associated with the persistent cookie if the persistent cookie persisted on the client system during the entire reporting period; determining a count per device value based on the weighted count data, the count per device value indicating an average number of beacon messages received per device during the reporting period; and determining a count of unique devices that have accessed the set of resources during the reporting period based on the count per device value and the total count of beacon messages received from the group of client systems.

12. The method of claim 11 wherein determining the persistent cookie data comprises: determining that a difference between a first time a cookie is received with a beacon message and a second time the cookie is received with a beacon message is greater than the threshold amount of time; and in response to determining that the difference is greater than the threshold, identifying the cookie as the persistent cookie.

13. The method of claim 12 wherein the reporting period is a month and the threshold amount of time is a threshold number of days.

14. The method of claim 11 wherein the reporting period is divided into subunits and the threshold amount of time is a threshold fraction of subunits in the reporting period, and determining the persistent cookie data comprises: determining that a fraction of the subunits in the reporting period in which a cookie is received with a beacon message exceeds the threshold fraction of subunits in the reporting period; and in response to determining that the fractions exceeds the threshold fraction of subunits in the reporting period, identifying the cookie as the persistent cookie.

15. The method of claim 11 wherein determining the weighted count data comprises: for a given persistent cookie: determining a total time present in the reporting period, wherein the total time present is a time in the reporting period that the given persistent cookie was received with a beacon message; determining a total time not present in the reporting period, wherein the total time not present is a time in the reporting period that the given persistent cookie was not received with a beacon message; determining a total number of times the given persistent cookie was received during the reporting period with a beacon message; and determining an estimated count of beacon messages that would have been received with the given persistent cookie from a client device associated with the persistent cookie if the persistent cookie persisted on the client device during the entire reporting period, based on the total time present, the total time not present, and the total number of times the given persistent cookie was received during the reporting period with a beacon message.

16. The method of claim 11 wherein determining the weighted count data comprises: for a given persistent cookie: determining an active persistency value, wherein the active persistency value is a percentage of the reporting period in which a beacon message with the given persistent cookie was received; determining an active subunit value, wherein the active subunit value is a count of subunits of the reporting period during which a beacon message with the given persistent cookie was received; determining a total time not present in the reporting period, wherein the total time not present is a time in the reporting period that the given persistent cookie was not received with a beacon message; determining a total number of times the given persistent cookie was received during the reporting period with a beacon message; and determining an estimated count of beacon messages that would have been received with the given persistent cookie from a client device associated with the persistent cookie if the persistent cookie persisted on the client device during the entire reporting period, based on the activity persistence value, the active subunit value, the total time not present, and the total number of times the given persistent cookie was received during the reporting period with a beacon message.

17. The method of claim 11 wherein determining the count per device value comprises determining the count per device value based on the estimated count of beacon messages and a total number of persistent cookies in the multiple persistent cookies.

18. The method of claim 11 comprising subdividing the persistent cookie data into one or more groupings, wherein each grouping corresponds to a category of client system, wherein: determining the weighted count data comprises determining the weighted count data for each of the groupings; determining the count per device value comprises determining the count per device value for each grouping based on the weighted count data for the grouping; and determining the count of unique devices comprises determining the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received from the group of client systems.

19. The method of claim 18 wherein determining the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received from the group of client systems comprises: subdividing the total count of beacon messages received from the group of client systems into the groupings; and determining the count of unique devices that have accessed the set of resources for each grouping based on the count per device value for the grouping and the total count of beacon messages received for the grouping.

20. The method of claim 18 wherein the groupings are based on type of client system, operating system of the client system, or provider of connectivity of the client system.