All case studies
Client under NDA

Demand forecasting for an FMCG distributor

A forecasting model and planning dashboard that cut stockouts and excess inventory at the same time.

Data & AnalyticsData Products & DashboardsLogistics, FMCGPoland

The challenge

The distributor managed around 4,500 SKUs with forecasts built in spreadsheets. Frequent stockouts on fast-moving products sat alongside excess stock of slow movers, tying up working capital and causing lost sales.

The question

Can we forecast demand accurately enough at SKU level to reduce both stockouts and excess inventory?

Our approach

  1. 01

    Data audit — three years of sales history, promotion calendars, pricing and supplier lead times.

  2. 02

    External drivers — holidays, seasonality, weather and promotional activity added as forecast inputs.

  3. 03

    Modelling — forecasting models tested by product group, with the best-performing approach selected for each.

  4. 04

    Pipeline — automated daily data refresh from the ERP into the forecasting engine.

  5. 05

    Planning tool — dashboard for buyers showing forecasts, recommended order quantities and exceptions.

What we delivered

  • SKU-level demand forecasting model with weekly forecasts 12 weeks ahead
  • Automated data pipeline from the client's ERP
  • Planning dashboard for the purchasing team
  • Handover, documentation and 3 months of support

Outcome

34% → 21%

Forecast error (MAPE) reduced from 34% to 21%

−28%

Stockouts on top-selling products down 28%

−15%

Excess inventory down 15%

Key insight

Most of the gain came not from the model itself but from putting promotions data into the forecast. Promotions explained a large share of the swings buyers had treated as unpredictable.

Related cases

Facing a similar decision? We'll walk you through relevant experience on a private call.

Book a private call