Krasnodar, Russia (April 19, 2023): Magnit, Russia’s largest retailer in terms of the number of stores, is rolling out the smart scale technology. The scales can automatically identify products being weighed by the customer, which accelerates the process and improves the quality of service.

To determine which product has been put on the platform, smart scales use a built-in camera to take pictures of the article. The photo is then sent to the image recognition server, where a neural network analyzes the picture and displays the best matches to the customer. By the end of 2023, the Company will integrate this technology into more than 1.500 self-service weighing stations in 200 of its supermarkets and superstores.

The system can identify the entire assortment of articles sold by weight, which includes over 800 SKUs in categories like fresh and frozen fruits and vegetables, sweets, ready-to-cook products, seafood, snacks, etc. The technology is intended to be used for products in transparent packaging (plastic or mesh bags). It boasts 98% recognition accuracy and is capable of self-learning by remembering the choices made by customers during weighing.

Magnit tested the smart scale technology in 37 of its large-format stores in 2021–2022. The pilot achieved a 0.3% improvement in profitability for categories sold by weight by reducing the number of instances when the products were misidentified. In 2023, the Company plans to implement a similar pilot project in its convenience stores, integrating the technology into the scales used by cashiers.

Eduard Iryshkov, Director of Retail Technology Department at Magnit:

“Introduction of smart scales enables us to improve the quality of our service thanks to the image recognition technology: now, customers do not have to remember the code of the product, and the weighing happens much faster. In addition to that, this solution helps us better manage the assortment and prevent misidentification of products, which is especially relevant for loose items. We are satisfied with the results of the pilot and ready to roll out the smart scale technology to 200 of our large-format stores.”

Magnit is already using image recognition powered by neural networks to monitor merchandising: artificial intelligence uses pictures of products on store shelves to check whether the items are placed in line with the pre-defined arrangement (also called a planogram). The system analyzes availability of products and their stocks in the store, order of on-shelf placement, product facings, and other parameters.

Magnit to Install Smart Scales in 200 Supermarkets and Superstores