DETERMINING TARGETING INFORMATION BASED ON A PREDICTIVE TARGETING MODEL

• US 20160078485A1
• Filed: 09/16/2014
• Published: 03/17/2016
• Est. Priority Date: 09/16/2014
• Status: Active Grant

• Current Assignee: Foursquare Labs, Inc.
• Original Assignee: Placed, Inc. (Foursquare Labs, Inc.)
• Inventors: Tryba, Jeremy, Rajaram, Siddharth, Grebeck, Michael, Yi, Weilie, Waldron, Elliott, Shim, David, Eatherly, Andrea, Gerner, Nick
• Granted Patent: US 10,423,983 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 30/0267 Wireless devices
• CPC Class Codes: G06Q 30/0271 Personalized advertisement
• CPC Class Codes: H04W 4/021 Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
• CPC Class Codes: H04W 4/029 Location-based management or tracking services

Abstract

A targeting system based on a predictive targeting model based on observed behavioral data including visit data, user profile and/or survey data, and geographic features associated with a geographic region. The predictive targeting model analyzes the observed behavioral data and the geographic features data to predict conversion rates for every cell in a square grid of predefined size on the geographic region. The conversion rate of a cell indicates a likelihood that any random user in that cell will perform a targeted behavior.

Claims

1. A method for use by at least one data processing device, the method comprising:

receiving targeting criteria, including a targeted behavior and a geographic region;

segmenting the geographic region using a grid into cells, wherein each cell has a cell identifier;

receiving behavioral information associated with multiple users, wherein the behavioral information includes time-stamped place visit data corresponding to visits to places by the multiple users;

calculating a behavior match metric for each cell based on the behavioral information;

receiving feature data for each cell;

labeling the feature data for each cell using the behavior match metric for the corresponding cell to obtain labeled feature data;

training a model for predicting a conversion rate of each cell based on a set of the labeled feature data, wherein the conversion rate provides a probability of a user in a cell performing the targeted behavior;

applying the model to the feature data to predict the conversion rate of each cell; and

identifying targeting information based on the conversion rates of the cells.

2. The method of claim 1, wherein identifying the targeting information based on the conversion rates of the cells further comprises:

sorting the cells based on the conversion rates; and

selecting a subset of the cells that have higher conversion rates compared to other cells.

3. The method of claim 2, further comprising providing latitude and longitude coordinates corresponding to the selected subset of the cells as the targeting information to an entity from which the targeting criteria were received.

4. The method of claim 2, further comprising:

determining one or more unique identifiers of users observed in the selected subset of the cells; and

,

providing the one or more unique identifiers as the targeting information to an entity from which the targeting criteria were received.

5. The method of claim 1, wherein identifying the targeting information based on the conversion rates of the cells further comprises:

calculating an aggregate conversion rate for each cell based on conversion rates of all cells within a predefined distance from the cell;

selecting cells with higher aggregate conversion rates than others; and

providing latitude and longitude coordinates corresponding to the selected cells as the targeting information to an entity from which the targeting criteria were received.

6. The method of claim 1, wherein the behavioral information further includes survey data and demographic profile data associated with the multiple users.

7. The method of claim 6, further comprising:

identifying a global user set from the multiple users based on at least one of the survey data or the demographic profile data;

wherein the global user set includes the multiple users when no criteria for identifying users based on the survey data or the demographic profile data is provided.

8. The method of claim 7, further comprising:

identifying a behavior match user set from the global user set based on the survey data;

wherein the behavior match user set is the same as the global user set when no criteria for identifying the behavior match user set based on the survey data is provided.

9. The method of claim 8, further comprising:

identifying a global visit set as a subset of all visits performed by users in the global user set; and

identifying a subset of the global visit set that matches the targeted behavior and was performed by a user in the behavior match user set as behavior match visit set.

10. The method of claim 6, wherein calculating the behavior match metric for each cell based on the behavioral information further comprises:

grouping the behavior match visit set by a user and by day;

for each user in the behavior match user set, calculating a total number of days on which the user performed the targeted behavior, wherein the total number of days corresponds to the behavior match metric.

11. The method of claim 1, wherein labeling the feature data for each cell includes using the cell identifier of the cell to join a key-value pair having a tuple of the cell identifier and a user identifier as a key and a behavior match metric of the cell as a value with another key-value pair having the cell identifier as a key and the feature data as a value.

12. The method of claim 1, further comprising:

generating the feature data for each cell, wherein the feature data includes;

nearby place features related to all places that are within a predefined distance from each cell;

place distance features related to minimum distances from each cell to each business identifier and each category identifier respectively that are within a predefined distance from the cell; and

demographic features for each cell obtained from census data.

13. The method of claim 1, wherein training the model for predicting the conversion rate of each cell based on the set of the labeled feature data further comprises:

training a classifier using the set of the labeled feature data to predict a visit probability that provides an indication of a likelihood that a user observed in the cell would perform the targeted behavior; and

training a statistical model using the visit probability and an observed conversion rate aggregated across all users in the cell to predict the conversion rate for the cell.

14. The method of claim 13, wherein applying the model to the feature data to predict the conversion rate of each cell further comprises:

applying the trained classifier on the feature data to predict a visit probability; and

applying the statistical model on the visit probability to predict the conversion rate for the cell.

15. The method of claim 14, further comprising:

randomly sampling users to obtain the set of the labeled feature data from the feature data for training the model.

16. The method of claim 1, wherein the targeted behavior includes visiting a place or performing an activity.

17. A system comprising:

memory;

at least one processor in communication with the memory and configured to execute a plurality of instructions stored in the memory to;

receive targeting criteria, including a targeted behavior and a geographic region;

segment the geographic region using a grid into cells, each cell having a cell identifier;

receive behavioral information associated with a plurality of users, the behavioral information including time stamped place visit data corresponding to visits to places by the plurality of users;

calculate a behavior match metric for each cell based on the behavioral information;

receive feature data for each cell;

label the feature data for each cell using the behavior match metric for the corresponding cell to obtain labeled feature data;

train a model for predicting a conversion rate of each cell based on a set of the labeled feature data, the conversion rate providing a probability of a user in a cell performing the targeted behavior;

apply the model to the feature data to predict the conversion rate of each cell; and

identify targeting information based on the conversion rates of the cells.

18. The system of claim 17, further configured to select a subset of the cells having higher conversion rates compared to other cells, wherein the targeting information includes at least one of:

latitude and longitude coordinates of the subset of the cells;

or

one or more unique identifiers associated with users observed in the subset of the cells.

19. A non-transitory computer-readable medium storing computer-executable instructions comprising:

instructions for gathering behavioral data, and sets of attributes associated with an array of locations in a geographic region; and

processing the gathered data to identify at least one of locations or users that are likely to convert to a particular campaign.

20. The medium of claim 19, further comprising:

instructions for providing promotional or non-promotional content associated with the particular campaign to other users based on the identified locations or the users.

21. The medium of claim 19, wherein the behavioral data includes one or more observed visit data, demographic profile data or survey data and each set of attributes is a high dimensional data generated based on place data and census data.