The evidence is in the working product and the choices behind it.
FMCG price intelligence
Decision context: comparing shelf prices across products and pack formats without confusing a change in pack size with a change in price.
Design: normalise observations to price per litre, preserve their history and make like-for-like comparisons explicit. The public dashboard separates regular prices, promotions and loyalty offers, compares Coca-Cola with Pepsi pack for pack, and benchmarks the full-sugar premium against Ireland’s sugar levy. Six sections cover the summary, sugar tax, pack architecture, promotions, competition and methodology.
Interpretation: observed prices support comparison and investigation. They do not, on their own, establish price elasticity, promotion ROI or causal effects. Coverage is one retailer; the levy ratio is a benchmark, not measured tax pass-through.
Inspect the price dashboard → · Read the current FMCG case study → · Read the implementation
Hierarchical demand forecasting
Decision context: demand planning needs forecasts that remain coherent across store, category and total levels.
Design: a 28-day forecast with reconciliation before publication. Evaluation considers the different business costs of stockouts and excess inventory.
Interpretation: forecast quality should be assessed in the context of the action it supports. This reference implementation does not establish a guaranteed commercial return.
Inspect the forecasting dashboard → · Read the implementation
Apply the pattern to your organisation
Start with a decision, identify the admissible evidence, make assumptions visible and decide how the outcome will be evaluated.