Delhaize lifts affiliated-store forecast accuracy about five percentage points on average across 600 stores

Delhaize, part of Ahold Delhaize Group, operates a pure B2B affiliate model, so it cannot push excess inventory to its store partners and the distribution center absorbs the cost when forecasts are wrong. Its ordering processes were reactive and manually intensive.

Delhaize deployed SymphonyAI's Demand Forecasting and Replenishment solutions across its 600-store affiliated network and five distribution centers in Belgium. The system generates daily order proposals at distribution center and store level, integrates with SAP Retail and accounts for vendor calendars, lead times, truck constraints, promotional pre-orders and available stock. Replenishment teams now work through exception-based review rather than validating every proposal, and affiliated store managers review only a handful of articles a day on handheld devices. SymphonyAI's Managed Services team supports the deployment.

At distribution center level, weekly forecast accuracy reaches 80-85% for dry and frozen categories and 85-90% in fresh, excluding promotional periods. For affiliated stores, accuracy has improved by approximately five percentage points on average and up to 10 percentage points for highly seasonal locations, and approximately 20% of DC order proposals are now auto-approved without user intervention. DC shrink has been decreasing across all departments without compromising service levels to affiliate partners, and working capital has improved materially.

• Exception-based review frees replenishment specialists to focus on critical alerts instead of validating every order proposal.
• In a pure B2B pull model with no inventory buffer, forecasting has to be right or the distribution center absorbs the cost.
• With SymphonyAI's Managed Services team, Delhaize reports forecast accuracy improving by 1.5 to 2 percentage points each year.