Demand intelligence.

A real-time, per-product model of shopper demand. Each product in the catalogue carries a continuously updated state — views, add-to-bags and purchases across nine time windows, plus lifecycle stage, momentum, category rank and stock. It answers “what is the current state of every product?” at any moment, rather than reporting on it after the fact.

Analytics is retrospective and aggregate: it tells you what happened, summarised across sessions, in dashboards you read later. Demand intelligence is live and per-product: it maintains the current state of each item, updated as events arrive, and is read by systems at request time. Analytics informs decisions people make later; demand intelligence feeds decisions systems make now.

Forecasting predicts future demand from historical patterns — typically on weekly or monthly horizons, for buying and supply decisions. Demand intelligence reads the present: momentum measured against each product’s own baseline, lifecycle transitions as they happen, scarcity as stock crosses a threshold. The two are complementary — but a forecast can’t tell you a product started surging twenty minutes ago.

Behavioural events from your storefront — views, add-to-bags, purchases — alongside inventory and catalogue data. Counters are maintained across nine rolling time windows per product, from the last few seconds to the last seven days. Momentum is measured against each product’s own recent baseline, not the catalogue average — so a small product accelerating registers, and a big product holding steady doesn’t.

No — social proof is one output of it. The same per-product model that selects storefront messages also drives the restock and markdown views in the portal and answers questions through Flockr Signal. Social proof messaging is demand intelligence rendered to shoppers; the model itself serves the whole stack.