The lost sale that nobody flags
A customer walks down aisle six looking for a specific brand of sparkling water. The shelf has been empty since the morning restock missed it, so the customer hesitates, checks the next aisle, and then leaves the store without buying. The transaction never appears in the day's sales report because the transaction never happened, and the shelf is still empty for the next customer and the one after that.
This pattern repeats across hundreds of SKUs every day in a typical large-format retailer. Global out-of-stock rates sit between five and ten percent of available sales depending on the category and the season, which means a chain doing one hundred million in annual sales is losing somewhere between five and ten million in revenue that walked through the door and walked back out without buying.
The cameras across the store see every empty shelf in real time, the store has already paid for them, and yet nobody is watching for empty shelves at the moment they appear because no one realistically can.
Why shelf checks are always too slow to catch the lost sale
Retailers have tried to fix out-of-stocks the same way they tried to fix theft, using manual shelf walks by associates on a rota, handheld scanners for store managers, and periodic audits by third-party shelf auditors. The model assumes someone is going to walk past the empty shelf and notice it within the brief window during which a restock would still capture the lost sale.
In a store with 25,000 SKUs across forty aisles, the rota model fails by design, because a shelf walk happens once or twice a shift and an empty shelf in between walks stays empty for hours. The customer who came in looking for the missing product has already left the store by the time the next walk reaches that aisle.
Inventory systems do not solve the problem either, because they tell the manager what was sold and what the system thinks is on the shelf, rather than what the customer is actually seeing when they look at the gondola in aisle six right now.
What changes when cameras start acting
Manako deploys Vision Agents directly onto the cameras a store already owns, with no new hardware to install, no engineers to hire, and no code for the team to write. The Vision Agent watches the shelves in each aisle continuously and writes a structured event the moment a designated facing drops below a defined threshold.
The event lands in the channels the store team already uses, including handheld devices, the team Slack /Whatsapp, or any other channel. The associate working that section receives a prompt with the aisle, the camera, and the timestamp, and walks over to check the shelf, restock from back stock, or trigger a replenishment from the warehouse before the customer has finished their loop of the store.
The same Vision Agent also flags repeat empty-shelf events on the same facing across the day, which gives the category manager a real signal about which SKUs are under-stocked relative to demand and which planograms need to change.
What restocked shelf hours are worth across the chain
The cameras and the in-store computing hardware already exist, which means Manako adds the Vision Agent layer that turns those cameras into a shelf-monitoring sensor without any new capital spend, any new infrastructure, or any additional staff on the rota.
A chain operating two hundred stores and losing an average of five percent of sales to out-of-stocks recovers a measurable share of that revenue within the first quarter of operation, and the recovery compounds across subsequent quarters as additional Vision Agents go live for additional categories, additional aisles, and additional planograms.
Retail vision is not only a loss prevention category
Retail vision has been treated as a loss prevention category for the last twenty years, with shrinkage detection, concealment alerts, exit fraud monitoring, and the post-incident review all working through the same camera infrastructure. The same cameras that catch a shoplifter are equally capable of catching the empty shelf that costs the business more revenue than the shoplifter ever did, because the underlying vision system is the same even when the specialist skills it runs are different.
Manako is the layer that selects the right skill for each situation and runs it on the cameras the retailer already owns.
Tell Manako what you need. It does the rest.
Join the waitlist to be the first to build with Manako, or get in touch to discuss an enterprise deployment across your operation.