Waste less with dynamic price point algorithm

As we are regularly informed, about one third of food produced globally is wasted. Supermarkets are a major cause as they have a lot of food with a limited expiry date. To reduce this particular source of food waste, the European Institute of Innovation & Technology (EIT) has supported Israeli start-up company Wasteless, which has developed a smart algorithm that allows the application of dynamic pricing markdowns based on expiry dates. 

“The aim is to encourage consumers to buy these products in time rather than letting them go to waste on the shelf,” explains David Kat, VP business development at Wasteless. “Applying discount stickers shortly before the expiry date is labour intensive and doesn’t entirely achieve the goal, as the food may start to look old or a bit devoured”, he explains. “The discount comes too late and is not efficient.”

“Supermarkets not only lose a lot of good food, but also a lot of money, 85% of the food waste is caused by fresh food that reaches its expiry date before it gets sold,” Kay states.

 Instead of waiting until the day before the expiry date to give a discount, retailer use the algorithm to calculate and test what the ideal date is to apply a discount. For example, a salad that expires on May 15 doesn’t need to have the same price as a salad that expires on May 21.

The reason they approach it this way is that consumers are very likely to prefer the salad with the later expiry date over the salad with the upcoming expiry date. “There is not necessarily a difference in taste or look, but the perceived value of the more recently made salad is simply higher”, Kat says. “Selling the older salad at a discounted price is a powerful tool to convince consumers to buy and use the product anyway.”

Stay informed: Get AIPIA News

The algorithm knows all the products on the shelves, including their expiry date and can predict whether or not it is likely that the product will get sold. Using these insights, it then calculates the most competitive price point for both consumers and the retailer. To evaluate and further improve the system, Wasteless uses reinforcement learning. This means that each time the algorithm decides a new price point, it will afterwards evaluate whether it was a correct analysis or if it should do it differently next time.

The algorithm can look at complementary products, Parmesan and Pasta for example, to make the discount even more interesting. The products have a data enabled barcode, which allows the supermarket to acquire data on the increased sales rate of products at those discounts. They can also add rewards or product information for consumers who buy sustainably.

« News feed