Shelf reality: when out-of-stock means out of sight
A brick-and-mortar chain needed to know what was on the shelf — and what wasn’t — without walking every aisle.
A brick-and-mortar chain needed to know what was on the shelf — and what wasn’t — without walking every aisle.
Store managers relied on manual counts and gut feel. Empty shelves were often discovered by customers, not staff. Planogram compliance was checked during audits, not continuously. HQ had no real-time view of on-shelf availability. They wanted to fix restock speed and compliance without adding headcount or invasive hardware in every store.
We deployed a vision-based system using existing in-store cameras where possible and low-cost nodes where needed. The system detects out-of-stock and low-fill conditions by aisle and shelf, compares layouts to planograms, and pushes alerts to store devices and a central dashboard. Alerts are prioritized so staff can tackle the biggest impact first. Data stays on-prem or in their cloud; we trained on their categories and planogram specs so the models speak their language.
On-shelf availability improved; restock cycles shortened because stores knew exactly where to go. Planogram compliance became measurable and improved over time. The chain is now trialling the same approach for queue and dwell analytics to optimize labor and layout.