Demand Sensing

Uses demand sensing on POS feeds to improve promotional and media-driven demand forecasting.

Business impact

• forecast accuracy — Danone reported forecast accuracy of 92% after a 20% reduction in forecast error.
• forecast error — The article reports a 20% reduction in forecast error.
• service levels — Service levels increased to 98.6% after a 30% reduction in lost sales.
• lost sales — The article reports a 30% reduction in lost sales.
• product obsolescence — The article reports a 30% reduction in product obsolescence.
• net uplift from promotions — The article reports a 55% improvement in net uplift from promotional events.

Data requirements

• point-of-sale feeds (tabular) — Used as input for demand sensing.
• promotional event information (tabular) — Promotions are part of the demand pattern the system models.
• media event information (tabular) — Media-driven demand is explicitly part of the forecasting problem.

AI methods and techniques

• demand sensing — Used to incorporate POS feeds from hypermarkets into forecasting.
• machine learning — Used to analyze demand variability and improve demand planning.

Industries

Real-world evidence

1 documented case study on record.

Companies using this: Danone.