Selected work
12Predictive models

FMCG Sales Forecasting

Hierarchical forecasting enhanced with external drivers, competitor effects, and machine-learning sub-models for stronger commercial planning.

Technology

AWSPython
Shopping cart filled with FMCG products in a supermarket aisle

Selected project note

01 / Context

The business context

In a fast-moving retail environment, accurate forecasting is critical for inventory planning, commercial decision-making, and operational performance. The forecasting landscape was complex, requiring models that could account for hierarchy, external drivers, and competitor effects.

02 / Solution

What was delivered

Enhanced hierarchical sales forecasting models by incorporating exogenous variables, competitor analysis inputs, and machine learning sub-models. Supported weekly-granularity forecasting with a one-year horizon and contributed to broader production, deployment, and version control processes associated with the forecasting pipeline.

03 / Impact

The practical impact

Delivered measurable forecast accuracy improvements that strengthened commercial planning, improved decision-making confidence, and contributed to more robust forecasting processes.