Tools for Tracking Your Customers and Measuring Shopper Engagement

Presentation TitleRaymond R. Burke and Alex Leykin
Kelley School of Business
Survey Research
Measure consumer perceptions of the shopping experience and diagnose problems with store, department, and category shoppability
Observational Research
Track shopper behavior, identify points of engagement and purchase obstacles, and
then manipulate and measure response
Copyright © 2007 Indiana University
Lead Fixtures and Merchandising
Category dwell time
Shopping basket size
24/7 tracking (time of day/crowding analysis)
Potential to track entire store (path analysis)
Scalable to multiple stores (benchmarking, experiments)
Speed:
Data Integration:
*
*
Disadvantages of RFID tracking:
Only applicable in retail stores using carts and/or baskets; e.g., grocery, mass retail
Only tracks customers who choose to use carts/baskets – loses fill-in shoppers
Unable to track customers who leave carts; may overestimate perimeter traffic, dwell times
No measure of gaze direction or package interaction
No information on group size or behavior
Limitations
Only applicable in retail stores using carts and/or baskets (e.g., grocery, mass retail)
Only tracks customers who choose to use carts/baskets, losing “fill-in” shoppers
Unable to track customers who leave carts. May overestimate perimeter traffic, dwell times
No measure of gaze direction or package interaction
No information on group size or behavior
*
Disadvantages of video tracking:
Cameras have a limited field of view and work best in smaller stores (e.g., specialty retail stores, drug stores, convenience stores, banks)
Occlusions (e.g., shelving, signage, other customers) can interfere with tracking
Difficult to distinguish between employees and customers
*
The GNC, Greeting Card areas of the store. Now, let’s take a brief look at what each of the cameras actually sees…
Chart1
Aisle1
Aisle2
Aisle3
MainAisle
LeftStoreRegion
Checkout Area 30%
0
0
0
0
0
0
0
0
0
0
Chart1
12am-1am
1am-2am
2am-3am
3am-4am
4am-5am
5am-6am
6am-7am
7am-8am
8am-9am
9am-10am
10am-11am
11am-12pm
12pm-1pm
1pm-2pm
2pm-3pm
3pm-4pm
4pm-5pm
5pm-6pm
6pm-7pm
7pm-8pm
8pm-9pm
9pm-10pm
10pm-11pm
28.4343434343
19.7272727273
13.9292929293
10.0252525253
9.1515151515
7.9595959596
13.904040404
53
76
87
75.0909090909
89.8434343434
105.0656565657
117.4696969697
119.8535353535
121.0202020202
107.6464646465
117.404040404
114.7777777778
108.5303030303
98.4747474747
86.7525252525
67.3131313131
EntranceCountsByTimeOfTheDay-We
Chart1
38415
38416
38417
38418
38419
38420
38421
38422
38423
38424
38425
38426
38427
38428
38429
38430
38431
38432
38433
38434
38435
38436
38437
38438
38439
38440
38441
38442
38443
38444
Count
Limitations
Cameras have a limited field of view and work best in smaller stores (e.g., specialty retail stores, drug stores, convenience stores, banks)
Tracking entire customer path requires multiple cameras with overlapping views
Occlusions (e.g., shelving, signage, other customers) and shadows can interfere with tracking
Difficult to distinguish between employees and customers
*
Each background pixel is represented as a stack of values
To decide if a new pixel is a part of the background, a lookup is performed through the full stack and if no matches are found the pixel is considered to be a “foreground pixel”
codebook
codeword
*
Each pixel in the image is modeled as a dynamically growing vector of codewords, a so-called codebook.
A codeword is represented by: the average pixel RGB value and by the luminance range Ilow and Ihi allowed for this
particular codeword. If an incoming pixel is within the luminance range and within some proximity of RGB of the
codeword it is considered to belong to the background.
During the model acquisition stage the values are added to the background model at each new frame if there is no match found in the already existing vector.
Otherwise the matching codeword is updated to account for the information from the new pixel. Empirically, we have established that there is seldom an overlap between the codewords.
The result of background subtraction is a binary bitmap
Foreground regions corresponding to moving people are represented as blobs (in red)
Tracking – Camera Model
Parallel lines and the heights of objects in the scene are used to determine the camera’s location and field-of-view
The camera model permits the translation from world coordinates to image coordinates and back
Tracking – Detecting Heads
The head is usually the least occluded part of the human body. Therefore, to reliably detect multiple people within one blob, we look at their head locations:
Estimate the height of each vertical line of the blob
Find a number of local maxima in the resulting histogram
Tracking – Detecting Heads (cont.)
Tracking – Probabilistic Modeling
At each instant in time, the tracking system attempts to find the model of the scene which:
Best fits the current observation (what’s in the image)
Is consistent with the model from the last observation
The system estimates the following parameters for each person:
body width and height (cm)
current location on the ground (X and Y)
color histogram
Tracking – Sampling Dynamics
To construct a new model, we randomly apply a number of “jump-diffuse” mutations to the old model
Then the likelihood of the new model is evaluated
Add body
Delete body
Store Entry
Shoppers take time and space to adjust to the in-store environment
Identify “recognition points” where consumers slow down and start observing
Provide answers and solutions, including signs, circulars, baskets, cash/wrap
Angle and direction of approach determines best position/orientation for signs and displays.
The greater the speed of approach, the
shorter the message
Facilitate incoming access to destination products, outgoing access to impulse items
Low penetration categories may require additional navigational aids, new product displays, merchandising, and/or changes in store layout to improve traffic flow
Categories with low purchase conversion rates may indicate weaknesses in product assortment, pricing, or presentation
*
Shirts and pants are folded in half to encourage product interaction
Enhanced product display drives category traffic and sales:
85% increase in product fixture interaction
44% increase in unit sales
38% increase in dollar sales
Reposition fixtures or product displays to eliminate bottlenecks
Avoid crowding in categories requiring extended decision times
Checkout
Measure queue lengths and waiting time to flag problems with line management, checkout process and customer service
Reduce waiting time by opening more lines, eliminating price checks, speeding up credit authorization, and employing self checkout
Total-traffic = Number of people (excluding children) who entered the store.
Left-traffic = Number of people (excluding children) who took the left path through the store.
Traffic
6-7AM
6-7AM
7-8AM
7-8AM
8-9AM
8-9AM
9-10AM
9-10AM
10-11AM
10-11AM
11-12PM
11-12PM
12-1PM
12-1PM
1-2PM
1-2PM
2-3PM
2-3PM
3-4PM
3-4PM
4-5PM
4-5PM
5-6PM
5-6PM
6-7PM
6-7PM
7-8PM
7-8PM
8-9PM
8-9PM
9-10PM
9-10PM
11/28/2003
11/21/2003
1421
0
726
0
709
0
719
0
798
83
795
100
602
136
620
118
592
112
438
122
423
140
388
182
351
209
317
250
298
204
124
87
Total-traffic = Number of people (excluding children) who entered the store.
Left-traffic = Number of people (excluding children) who took the left path through the store.
Traffic
6-7AM
6-7AM
7-8AM
7-8AM
8-9AM
8-9AM
9-10AM
9-10AM
10-11AM
10-11AM
11-12PM
11-12PM
12-1PM
12-1PM
1-2PM
1-2PM
2-3PM
2-3PM
3-4PM
3-4PM
4-5PM
4-5PM
5-6PM
5-6PM
6-7PM
6-7PM
7-8PM
7-8PM
8-9PM
8-9PM
9-10PM
9-10PM
11/28/2003
11/21/2003
0.1808585503
0
0.3567493113
0
0.4442877292
0
0.438108484
0
0.335839599
0.3734939759
0.3081761006
0.5
0.3122923588
0.2941176471
0.2548387097
0.3559322034
0.2533783784
0.3839285714
0.3401826484
0.393442623
0.3073286052
0.3
0.2628865979
0.3241758242
0.2849002849
0.3492822967
0.2839116719
0.296
0.2483221477
0.3431372549
0.4193548387
0.3563218391
Measure line queues and crowding
Cluster shoppers based on path similarity
Evaluate store layout and product adjacencies
Manage in-store communication, product assortment, and pricing
Manage service levels, staffing
www.kelley.iu.edu
0
20
40
60
80
100
120
140
12am-1am
1am-2am
2am-3am
3am-4am
4am-5am
5am-6am
6am-7am
7am-8am
8am-9am
9am-10am
10am-11am
11am-12pm
12pm-1pm
1pm-2pm
2pm-3pm
3pm-4pm
4pm-5pm
5pm-6pm
6pm-7pm
7pm-8pm
8pm-9pm
9pm-10pm
10pm-11pm
0
200
400
600
800
1000
1200
1400
1600
6-7AM7-8AM8-9AM9-10AM10-
11AM
11-
12PM
12-1PM1-2PM2-3PM3-4PM4-5PM5-6PM6-7PM7-8PM8-9PM9-10PM
11/28/2003
11/21/2003
0%
10%
20%
30%
40%
50%
60%
6-7AM7-8AM8-9AM9-10AM10-
11AM
11-
12PM
12-1PM1-2PM2-3PM3-4PM4-5PM5-6PM6-7PM7-8PM8-9PM9-10PM
11/28/2003
11/21/2003

Documents

Tools for Tracking Your Customers and Measuring Shopper Engagement