.png)
Ahold Delhaize
Retail · Europe
250 tonnes less food waste
Ahold Delhaize uses AI forecasting and dynamic markdowns to cut food waste by 250 tonnes
Forecasting model 26% more accurate. Dynamic markdowns every 15 minutes.
Friction
Food retail has structural tension between availability and waste. Margin pressure, sustainability obligations, and the need for scalable solutions across brands and countries.
Breakthrough
Two complementary levers that reinforce each other: better forecasting to buy less excess, and faster granular markdowns to lose less when overstock occurs anyway.
Impact
Delhaize Belgium: forecasting model 26% more accurate, 21% food waste reduction potential. Albert Heijn NL: dynamic markdowns every 15 minutes, 250+ tonnes less waste in year one.
Unlock the full analysis with breakthrough, impact, what made it smart and its technical approach below!
Problem
Food retail faces structural tension between product availability and waste. Margin pressure, sustainability obligations, and the need for solutions that scale across brands and countries.
What made it smart
Two mutually reinforcing levers in one programme. Forecast better to buy less excess, and mark down faster to lose less when overstock happens anyway. Each makes the other more effective.
Technical approach
Forecasting uses multiple data sources (promotions, time of day, weather) to predict demand. Dynamic markdowns translate inventory, shelf life, and demand forecast into frequent price updates, made operationally possible by electronic shelf labels and digital pricing processes.
Strategic lesson
Waste reduction at scale requires both prediction and action. Better forecasting alone is not enough if the operational response is too slow.
Reflection question
Where does your organisation systematically over-produce, over-buy, or over-staff, because you are optimising for availability rather than accuracy?