28 Dec 2017

Shopper experience analytics


A study on video analytics tools in retail

Colgram, A large retail chain based in Chile, had been using infrared beam technology to count shoppers entering or shoppers exiting their stores, but the retailer found that the data that was captured was not accurate, not comprehensive or not timely.
“We wanted a solution that would enable us to move beyond the door and analyze shopper behavior throughout the full store,” says Hernán Ferragut, director of retail for Colgram. The newer “sophisticated real-time data capture capabilities, combined with granular analytics and reporting, will give us immediate access to accurate information about store traffic, allow us to count children and adults as a single shopping unit and understand other key metrics,” says Ferragut.
Video surveillance is a mainstay in the retail industry. For years it’s been a vital tool, aiding retailers in security and loss prevention efforts. But retailers know the value of surveillance footage is not limited to ordinary security applications, so they are continuously pursuing innovative ways to turn raw video into “video-based data.” The development of intelligent video processing algorithms is bringing many new applications for video within stores, both in the traditional domain of loss prevention, and in store operations and merchandising.

Customer counting

The number of customers entering a store is one of the most important statistics of interest to a store manager that cannot be obtained from the data. Several specialized methods for counting customers are available (beam breakers, pressure pads, or staff with click counters) but all have drawbacks in accuracy or expense. In this system customer counting is carried using video analytics to count people passing through known entry or exit points. The entrance and exit counts can be used in conjunction to estimate the number of customers in the store at any time and the average peak time in the store.

Who are my customers?

Retailers need to know the demographic makeup of their customers to make better decisions about advertising, product placement, and merchandise mix. By parsing video streams, video analytics solutions can identify shoppers by age and gender, providing retailers with valuable statistical information about their potential customers.

How many people are in my store, and where do they go once they’re inside?

Almost as important as counting the number of people in the store knows where people go within the store. To this end the interface includes a department or camera level traffic analysis tool. This graphs the traffic in a particular camera over time, and can also visualize results both by drawing the tracks themselves or a “heat map” showing areas with greatest activity.

Is my marketing strategy effective? Where? Why?

Display effectiveness is evaluated in the interface by calculating statistics on where customers spend their time in a view of a display. The interface allows the user to choose active regions, such as the area in front of a display, and observe how many customers entered the region in a period of time, how many stopped there and how long these customers spent. All the trajectories of customers are shown and allow the user to “drill down” to the original video to observe the behavior of the selected customers.

What’s my store’s conversion rate?

The store wide counts can also be used to calculate average sales per customer or “conversion rate”. The number and value of sales in each hour can be calculated from the data. Dividing by the number of customers leaving the store (leaving times will correlate more closely with sales than entrance times), number and value of sales per person entering can be calculated, and graphed in the interface. Local people counts can also give estimates of conversion rates per store or per display.

Sales per square foot

is now

Experience per square foot

TetherBox Technologies
2nd Floor, P.V.R Complex, 80 Feet Road, IndiraNagar Bengaluru, Karnataka 560038 IN
Phone: 08025255133 Website: http://myklik.io