Datum · Retail
The shelf tells youbefore it's empty.
Datum is Nextrion's retail deployment.
Inventory, foot traffic, and promotions resolve into one graph per store, so restock and display decisions happen in the moment, not in Monday's report.
What’s connected
Every surface writes into one model.
D · 01POSSales velocity per SKU, per hour, as live entities.
D · 02InventoryWarehouse counts and shelf state in one graph.
D · 03VisionShelf gaps and foot traffic, anonymized.
D · 04PromotionsWhat's running, where, and what it actually moves.
D · 05StaffingFloor coverage matched to traffic, not habit.
D · 06ChannelsRestock triggers land in Slack, not a dashboard.
How it works
01
The store is modeled
Sales, stock, and movement land in one typed graph per store. Entity resolution matches the same SKU across POS, warehouse, and shelf systems that never agreed.
02
The gap is caught
The ontology relates shelf state to sales velocity and traffic, so an empty facing is a detected event, not a Monday discovery.
03
The move is made
Restock triggers, display swaps, and staffing shifts leave the model as actions: a task in Slack, an updated plan, a draft order.
Every store runs like your best one walked the floor.