Can intraday discounts save fresh products before they become waste?

Challenge

A grocery retailer was looking to reduce waste in fruit and vegetables sold in bulk, while preserving product freshness, costumer experience and store margin. Unlike packaged products, these items do not carry an individual expiry date. Instead, each batch has a freshness window that determines how long it can remain in store, making it harder to identify stock at risk and decide when a discount should be applied.

The challenge was to study the possibility of introducing a discount during the day that would help sell stock at risk before it had to be removed from the store, without unnecessarily discounting products that could still be sold at full price.

‍

Solution

A decision-support model was developed to simulate, throughout the day, which products were at risk of waste and whether a markdown, applied at a given hour, would generate a positive business impact. The model combined store-level sales forecasts, intraday sales patterns, available stock and historical promotional response to simulate different discount levels and application times.

For each day, product and store, it estimated the expected trade-off between additional sales, avoided waste, promotional cost, operational effort and potential margin loss from discounting stock that could otherwise be sold at full price. This approach made it possible to identify the best time to apply the intraday discount, as well as the most appropriate discount level depending on the product and store.

‍

Results

The study confirmed that intraday markdowns could help reduce waste while protecting margin.

The optimal scenario was identified as applying the discount around mid-afternoon, with 4 p.m. delivering the strongest impact overall.

Compared with a no-discount baseline, this scenario increased sales volume by around 2.7% and sales value by around 3.1% while keeping front-office margin value broadly stable, with a reduction of approximately 0.71 p.p.

The analysis also showed that discount timing is critical: earlier discounts were more likely to erode margin by discounting stock that could still be sold at full price, while later discounts had less time to generate enough additional sales before product removal. Given the positive expected impact and operational considerations involved, the recommended next step was to launch a pilot to validate the approach in real-store conditions.

‍