NewExplore decision agents and merchant-set guardrails
See how they workDecision chain · 01
Retail demand forecasting estimates how many units customers are likely to buy for each product, location and period, together with the uncertainty around that estimate.
Last reviewed
A planning forecast needs more than one headline number. It should identify the product and location, state the time bucket, show a plausible range and preserve the drivers that changed the result.
That makes the output usable by a merchant: the forecast can be challenged, approved and passed into the next decision instead of remaining a chart.
The forecast becomes an input to the merchandise plan, assortment, allocation and later pricing decisions. Keeping those decisions connected avoids re-keying and makes variances traceable.
Typical inputs include dated sales, inventory positions, product attributes, locations, prices, promotions, receipts and a retail calendar. The exact minimum depends on the decision being supported.
Allocation and replenishment happen at product-location level. A forecast at a broader level must be disaggregated before anyone can act on it, which can hide local size, store and channel differences.