How retailers can eliminate checkout, like Amazon Go. And increase profits in the process.

Amazon is shaking the world of retail with their checkout-free store. It is a revolutionary shopping experience. No lines, no checkout: just grab and go.

Eliminating checkout is an old dream of the retail industry. However, the Amazon solution is new. It mostly involves computer vision: Amazon Go uses cameras disseminated all around the store, with real-time video analysis by artificial intelligence. The previous attempt to eliminate checkout, 10 years ago by IBM, required sticking RFID tags on all products, an expensive and impractical process for store suppliers.

Most retailers are not prepared to address this new disruption. For example, Monoprix, a major French retailer, merely reacted with a parody of the Amazon Go clip:

All French shoppers love the legendary Monoprix sense of humor, but funny jokes are not enough to beat the Amazon Go customer experience. Monoprix must take artificial intelligence more seriously. Otherwise, they will disappear like Prisunic, or if they are more lucky, they might be acquired by their Seattle competitor, and become a kind of ‘Amazonoprix’.

The good news for Monoprix is that no huge upfront R&D investment is necessary to catch up with Amazon Go. Retailers can eliminate checkout by progressively increasing their use of AI in video. Moreover, each step improves their revenue and customer experience.

For example, stores can use AI cameras for merchandising placement, shelf monitoring, and shoplifting detection. This will prepare them for a smooth checkout-free transition. Let me develop:

Merchandising placement

Heatmap: in red the areas where consumers spent the most time.

In-store cameras can be used to optimize store layout. By tracking zones of customer movement with heatmaps, retailers can identify products that are visible to the customers, and those lacking their attention. Therefore, product pricing and placement can be adapted.

This technology only tracks movements, it doesn’t identify people, and preserves customers anonymity.

Value: optimizing product placement can increase sales by 5 to 20 percent.

Shelf monitoring

Object recognition on a store shelf

The second step in the checkout-free transformation is shelf monitoring. In a store, it often happens that items are misplaced, or shelves are empty. This inventory distortion is costly. Traditionally, visual check of shelves is performed by human employees, a task known as planogram compliance monitoring. Using object recognition technologies, in-store cameras can automatically recognize items on shelves, and assist humans in this tedious process.

Automatic item recognition also helps for omnichannel retail strategies: with real-time shelf monitoring, the store can be used as a warehouse for online orders. Moreover, by snapping a picture of the product of interest, customers can learn more about it online, on their smartphone.

Value: inventory distortion costs globally nearly $1.1 Trillion. On average, each store could increase sales by 7.5 percent, if it could completely fix this problem.

Shoplifting detection

Shoplifter stealing Oreo cookies

The next step in the checkout-free transition is to detect shoplifting. For example, by combining item tracking with another AI method called action recognition, the computer can automatically spot shoplifting acts on the surveillance camera, without RFID.

Value: shoplifting costs US retailers more than $16 Billion per year.

Towards checkout-free stores

The last step is to make shoplifting itself a thing of the past, by switching to a checkout-free experience. To achieve this last step, it remains to identify customers when they pick items on the shelves. Customers can be localized with their smartphone GPS, and further identified with face recognition and faceless recognition. In the situation when two different customers are located almost at the same place, the customer picking the item can be identified with hand skin color recognition.

Conclusion

A checkout-free store seems out of reach for most retailers. It is a radical transformation requiring full mastery of the latest AI technologies. However, this goal can be split into intermediate steps, which are much easier, yet profitable. In this process, retailers can gather experience, collect data, and get prepared to rival with the Amazon Go solution.

Checkout-free stores are not the exclusive privilege of technological giants like Amazon. Traditional retailers, including Monoprix, have the potential to develop their own in-house solution, or in partnership with an AI startup like Mindolia.

Check out our video analytics solutions for retail at mindolia.com/retail