With a worldwide call for bids, Aldi Süd is looking for start-ups in the fields of machine learning and computer vision that could help the discounter to easily skip generations of technology evolution.
On a simple registration page, the discounter asks young entrepreneurs to introduce themselves when they have developed solutions ready for testing that enable the automatic recognition of articles: on shelves through cameras, through apps on consumers’ smartphones or even in shopping carts. And all of this on the basis of computer vision.
Aldi Süd explicitly excludes solution providers that rely on RFID. That is clever. Equipping every yoghurt with a microchip is neither economically nor ecologically sensible.
The automatic recognition of articles by cameras is a basic technology that could radically optimise most in-store processes in the 6,520 Aldi Süd stores worldwide. And eliminate the two biggest annoyances for customers: Out-of-Stocks and queues at the checkout.
Additional product information on shoppers’ apps
However, Aldi Süd also explicitly mentions shopper apps as one area of application, which display additional information on the products to the customers. Numerous retail companies are currently looking for such solutions in order to make up for the lead of online trade in the filtering of product attributes. People who have several allergies or special nutritional needs tend to buy their food online because finding suitable products in brick and mortar retail currently represents a major challenge for them.
However, Aldi Süd’s tender is about much more than additional product information for customers: Cameras could monitor at any time nonstop whether the actual presence of goods on the shelves corresponds to the planograms. Employees should be alerted — not only in the case of out-of-stocks, but also in the case of cold de-listings, where not only the goods but also the shelf price label is missing.
Dream of fully automated stock taking
While the facings of the shelves can easily be captured by cameras, the far greater challenge is to automate the entire stock taking process. Sensors in the shelves can play a role here. Without a doubt, the increase in efficiency and the cost reduction would be gigantic for Aldi if the discounter always had the correct stock in the merchandise management system in real time, fully automated.
Last but not least, Aldi Süd is apparently planning to skip the technological evolutionary stages of self-checkout and self-scanning and is looking for solutions that automatically recognize which items shoppers place in the trolley or which products are in it when it comes to payment.
It will be interesting to see whether codes integrated into the packaging design play a role here, as Edeka’s Netto is implementing them with Digimarc to speed up the scanning process at the checkout. The decisive advantage of these solutions is that the article number can be read by machine from any side of the product, even if the majority of the article is concealed.
It is still unclear, however, whether the packaging design needs special coding at all, or whether it will be possible to identify articles without error even on the basis of their usual packaging design: and this with thousands of articles standing close together on the shelf.