Brand-switching analysis using longitudinal tracking of at-shelf shopper behavior

• US 10,354,262 B1
• Filed: 06/02/2016
• Issued: 07/16/2019
• Est. Priority Date: 06/02/2016
• Status: Active Grant

• Current Assignee: Videomining Corporation
• Original Assignee: Videomining Corporation
• Inventors: Hershey, Jeff, Sharma, Rajeev, Hirata, Richard, Botla, Krishna
• Primary Examiner(s): Waesco, Joseph M
• Application Number:
  US15/171,873
• Time in Patent Office: 1,139 Days
• Field of Search: None
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

A method and system for analyzing shopper switching behavior over multiple shopping trips. Specifically, shopper behavior can be analyzed to generate a shopper behavior profile. Aggregating profile data across many shopping trips over time can then be used to generate analytics regarding the shopper'"'"'s behavior regarding switching between products or brands and whether the shopper'"'"'s purchase decisions occurred at-shelf or pre-shelf. Further, the data can be aggregated across many shoppers The analysis can then be used to make recommendations to manufacturers or retailers about how to influence shoppers to switch to a brand from a competitor, or to prevent a shopper from switching to a competitor. A deployment of cameras and mobile signal sensors can be utilized to recognize shoppers and track their behavior. Demographics information can also be estimated about the tracked shoppers. The visual and mobile signal trajectories can be fused to form a single shopper trajectory, then associated with Point of Sale (PoS) data. This results in a dataset describing the shopping trip for each tracked shopper.

Claims

1. A method for performing brand switching analysis by determining behavior for at least one shopper during a plurality of shopping trips, for at least one retail location, utilizing at least a sensor, and at least a processor for performing the steps of:

a. tracking behavior of a shopper at a retail location, wherein tracking the behavior of a shopper further comprises;

i. obtaining a set of vision data from at a camera,

ii. detecting a shopper at a specific time and location,

iii. obtaining a set of mobile data for the shopper using a mobile device,

iv. localizing the mobile device using the MAC address using a trilateration based method,

v. tracking the shopper using the mobile device using the set of vision data and the set of mobile data through a Multi-modal Trajectory Fusion module,

vi. creating a set of shopper profile data using the shopper trajectory,

b. integrating a set of data from the Multi-modal Trajectory Fusion module using a Multi-modal Shopper Data Associator, which comprises the following steps;

i. detecting the completion of at least one mobile trajectory,

ii. retrieving a set of shopper profile data from the in-store shopper database, wherein the shopper profile data contains at least one vision trajectory,

iii. performing matching between the at least one vision trajectory and the at least one mobile trajectory,

iv. fusing vision trajectories that are associated with the same target at a given time frame using measurement fusion,

v. combining the fused vision trajectories with the mobile trajectory to complete missing segments in the vision trajectories,

c. calculating at least one decision factor using a Shopper Decision Tracker module,

d. determining whether a shopper decision was made at-shelf or pre-shelf, based on the at least one decision factor, using a Decision Determination module,

e. recognizing a returning shopper to a tracked location, or to a different location, and aggregating data from multiple trips by the shopper to a shopper database,

f. creating a profile for the shopper, based on the shopper'"'"'s behavior during the plurality of shopping trips, and storing the profile in the shopper database,

g. calculating derivable metrics and adding the derivable metrics to the shopper profile, using a Shopper Behavior Profiler module,

h. analyzing the shopper profile to determine the cause for brand switching that has occurred during the time the shopper behavior was tracked, using the Shopper Behavior Profiler module, and

i. using an Analytics Generation module for one or more of;

i. analyzing the shopper profile to make recommendations to a retailer or brand manufacturer for influencing the shopper to switch to their brand from a competitor,

ii. analyzing the shopper profile to make recommendations to a retailer or brand manufacturer for influencing the shopper to not switch from their brand to a competitor'"'"'s brand, and

iii. analyzing shopper profiles, aggregated across a plurality of shoppers, to derive metrics representing the strength of a product or brand of products.

2. The method of claim 1, wherein the steps are performed for a plurality of shoppers at a single retail location or across multiple retail locations, and the resulting calculated data are aggregated in the shopper database.

3. The method of claim 1, wherein the Shopper Decision Tracker module uses one or more factors from a list comprising time based factors, category based factors, environmental factors, and behavior based factors and wherein the list of time based factors is comprised of time allocation, shopping time, navigation time, and navigational complexity, the list of category based factors is comprised of category size, category orientation, category complexity, fixture type, and share of category space, the list of environmental factors is comprised of frequently changing elements, aisle fixtures, and crowding, and the list of behavior based factors is comprised of stopping frequency, shopper velocity, shopper trajectory, and shopper repeat visits.

4. The method of claim 1, wherein the Shopper Decision Tracker module utilizes at least one decision factor to populate a decision model used by the Decision Determination module, wherein the decision model is generated by the steps of:

a. performing an association analysis for identifying factors that have a bivariate association with one another using an Association Analysis module,

b. estimating the probability that, given a set of factors, that a particular outcome is present, using a Multiple Logistic Regression module, and

c. scoring the model for goodness of fit, applying necessary transformations based on the scoring, and refining the model using a Model Validation and Optimization module.

5. The method of claim 1, wherein the derivable metrics comprise at least one of a Brand Switching Event Timeline, a Decision Behavior History, a Brand/Product Loyalty Estimate, a Brand/Product Affinity Index, a Brand/Product Loyalty Cluster, a Brand/Product Affinity Cluster, a Store Loyalty Estimate, a Household Income Estimate, a Shopping Style Classifier, a Household Size Estimate, a Shopper Channel Distribution, a Healthy Food Preference Estimate, a Shopper Channel Preference, an Organic Food Preference Estimate, a Sensitivity Rating, a Shopper Switching Profile, and a Stimulus Effectiveness Rating.

6. The method of claim 5, wherein the Brand Switching Event Timeline further comprises at least one of:

a. matching at least one stimulus to at least one change in decision behavior, and

b. matching at least one stimulus to at least one brand switching event.

7. The method of claim 5, wherein the Decision Behavior History comprises data indicating whether the shopper decided at-shelf or pre-shelf for every purchase made by the shopper, wherein the Sensitivity Rating indicates the degree stimulus needed to affect shopper behavior, wherein the Shopper Switching Profile identifies switching patterns within or across brands, and wherein the Stimulus Effectiveness Rating indicates how effective a particular stimulus is in causing a change in shopper behavior.

8. The method of claim 1, wherein the Analytics Generation module comprises a Shopper Level Switching Analysis module and a Brand Level Switching Analysis module, wherein the Shopper Level Switching Analysis module generates a summary of the shopper'"'"'s switching behavior over time, and wherein the Brand Level Switching Analysis module generates a summary of switching patterns for a brand or product over time.

9. The method of claim 8, wherein the Shopper Level Switching Analysis module predicts shopper switching behavior based on the shopper'"'"'s behavior history.

10. The method of claim 8, wherein the Analytics Generation module further comprises providing an interpretation of the calculated metrics and making recommendations for improving the strength of a product or brand, retaining customers, or enticing shoppers to switch from a competitor, using a Brand Switching Scorecard.

11. A system for performing brand switching analysis by determining behavior for at least one shopper during a plurality of shopping trips, for at least one retail location, utilizing at least a sensor, and at least a processor for performing the steps of:

a. tracking behavior of a shopper at a retail location, wherein tracking the behavior of a shopper further comprises;

i. obtaining a set of vision data from at a camera,

ii. detecting a shopper at a specific time and location,

iii. obtaining a set of mobile data for the shopper using a mobile device,

iv. localizing the mobile device using the MAC address using a trilateration based method,

v. tracking the shopper using the mobile device using the set of vision data and the set of mobile data through a Multi-modal Trajectory Fusion module,

vi. creating a set of shopper profile data using the shopper trajectory,

b. integrating a set of data from the Multi-modal Trajectory Fusion module using a Multi-modal Shopper Data Associator, which comprises the following steps;

i. detecting the completion of at least one mobile trajectory,

ii. retrieving a set of shopper profile data from the in-store shopper database, wherein the shopper profile data contains at least one vision trajectory,

iii. performing matching between the at least one vision trajectory and the at least one mobile trajectory,

iv. fusing vision trajectories that are associated with the same target at a given time frame using measurement fusion,

v. combining the fused vision trajectories with the mobile trajectory to complete missing segments in the vision trajectories,

c. calculating at least one decision factor using a Shopper Decision Tracker module,

d. determining whether a shopper decision was made at-shelf or pre-shelf, based on the at least one decision factor, using a Decision Determination module,

e. recognizing a returning shopper to a tracked location, or to a different location, and aggregating data from multiple trips by the shopper to a shopper database,

f. creating a profile for the shopper, based on the shopper'"'"'s behavior during the plurality of shopping trips, and storing the profile in the shopper database,

g. calculating derivable metrics and adding the derivable metrics to the shopper profile, using a Shopper Behavior Profiler module,

h. analyzing the shopper profile to determine the cause for brand switching that has occurred during the time the shopper behavior was tracked, using the Shopper Behavior Profiler module, and

i. using an Analytics Generation module for one or more of;

i. analyzing the shopper profile to make recommendations to a retailer or brand manufacturer for influencing the shopper to switch to their brand from a competitor,

ii. analyzing the shopper profile to make recommendations to a retailer or brand manufacturer for influencing the shopper to not switch from their brand to a competitor'"'"'s brand, and

iii. analyzing shopper profiles, aggregated across a plurality of shoppers, to derive metrics representing the strength of a product or brand of products.

12. The system of claim 11, wherein the steps are performed for a plurality of shoppers at a single retail location or across multiple retail locations, and the resulting calculated data are aggregated in the shopper database.

13. The system of claim 11, wherein the Shopper Decision Tracker module uses one or more factors from a list comprising time based factors, category based factors, environmental factors, and behavior based factors and wherein the list of time based factors is comprised of time allocation, shopping time, navigation time, and navigational complexity, the list of category based factors is comprised of category size, category orientation, category complexity, fixture type, and share of category space, the list of environmental factors is comprised of frequently changing elements, aisle fixtures, and crowding, and the list of behavior based factors is comprised of stopping frequency, shopper velocity, shopper trajectory, and shopper repeat visits.

14. The system of claim 11, wherein the Shopper Decision Tracker module utilizes at least one decision factor to populate a decision model used by the Decision Determination module, wherein the decision model is generated by the steps of:

a. performing an association analysis for identifying factors that have a bivariate association with one another using an Association Analysis module,

b. estimating the probability that, given a set of factors, that a particular outcome is present, using a Multiple Logistic Regression module, and

c. scoring the model for goodness of fit, applying necessary transformations based on the scoring, and refining the model using a Model Validation and Optimization module.

15. The system of claim 11, wherein the derivable metrics comprise at least one of a Brand Switching Event Timeline, a Decision Behavior History, a Brand/Product Loyalty Estimate, a Brand/Product Affinity Index, a Brand/Product Loyalty Cluster, a Brand/Product Affinity Cluster, a Store Loyalty Estimate, a Household Income Estimate, a Shopping Style Classifier, a Household Size Estimate, a Shopper Channel Distribution, a Healthy Food Preference Estimate, a Shopper Channel Preference, an Organic Food Preference Estimate, a Sensitivity Rating, a Shopper Switching Profile, and a Stimulus Effectiveness Rating.

16. The system of claim 15, wherein the Brand Switching Event Timeline further comprises at least one of:

a. matching at least one stimulus to at least one change in decision behavior, and

b. matching at least one stimulus to at least one brand switching event.

17. The system of claim 15, wherein the Decision Behavior History comprises data indicating whether the shopper decided at-shelf or pre-shelf for every purchase made by the shopper, wherein the Sensitivity Rating indicates the degree stimulus needed to affect shopper behavior, wherein the Shopper Switching Profile identifies switching patterns within or across brands, and wherein the Stimulus Effectiveness Rating indicates how effective a particular stimulus is in causing a change in shopper behavior.

18. The system of claim 11, wherein the Analytics Generation module comprises a Shopper Level Switching Analysis module and a Brand Level Switching Analysis module, wherein the Shopper Level Switching Analysis module generates a summary of the shopper'"'"'s switching behavior over time, and wherein the Brand Level Switching Analysis module generates a summary of switching patterns for a brand or product over time.

19. The system of claim 18, wherein the Shopper Level Switching Analysis module predicts shopper switching behavior based on the shopper'"'"'s behavior history.

20. The system of claim 18, wherein the Analytics Generation module further comprises providing an interpretation of the calculated metrics and making recommendations for improving the strength of a product or brand, retaining customers, or enticing shoppers to switch from a competitor, using a Brand Switching Scorecard.