Empty Shelf and Out-of-Stock (OOS) Detection
Out-of-stock detection turns the camera in the aisle into a shelf check: it flags the shelf facing that has emptied and sends an alert to the aisle manager's device. ALGI runs this check on your existing store cameras, with no shelf sensors, smart shelves or extra hardware. A product that is not on the shelf cannot be sold, yet the lost revenue often never shows up in the stock records, because the product is sitting in the back room. The system makes that gap visible: which shelf, when, and for how long it stayed empty.
What is out-of-stock (OOS) detection?
Out-of-stock detection is the identification of gaps on the shelf from camera video. The system treats shelf facings as zones and flags when product fill in a zone drops below a defined threshold.
The problem cannot be solved from stock records. A shelf that is empty while stock sits in the back room, whether because of misplacement, a late replenishment run or a drifted planogram, never appears in the inventory system. A measurement that looks at the shelf itself shows that gap directly.
The measurement is anonymous. The system looks at the shelf, not at shoppers; it does not count or identify people. The output is a shelf zone, a time and the duration the zone stayed empty. The same data also supports planogram compliance: it compares whether the product group that belongs on a shelf is actually in place.
Because every alert carries the shelf, the time and the duration, the log becomes a record of how replenishment really runs on the floor, shift by shift and aisle by aisle. That record is what lets store and category teams agree on where the real gaps are.
How does out-of-stock detection work?
The check is set up in three steps.
Draw the shelf zones. In the dashboard, shelf facings or product groups are marked as zones over the camera view. A separate fill threshold that counts as empty is defined for each zone, because boxed products and bagged products are not measured against the same threshold.
Process the video. The camera stream is processed either on an Algi Edge device installed on site or on ALGI AI servers, and shelf fill is assessed at regular intervals. With an on-site Algi Edge device, live video stays in the store and only fill data tied to zone and time leaves the site.
Deliver the alert to the aisle. When a zone drops below its threshold and stays there for a defined period, an alert opens and a mobile notification goes to the aisle manager. A waiting period is configured so that momentary fluctuations do not turn into alerts.
Records are reported by shelf, hour and day. The report shows which shelf regularly empties at which hour, and the replenishment plan changes accordingly. Read together with density data, it also shows which sales hour an empty shelf coincided with.
Which industries use it?
Frequently asked questions
- What is out-of-stock detection and how does it work?
- Out-of-stock detection assesses product fill in the shelf zones you draw in the dashboard using camera video. When fill drops below a defined threshold and stays there, it opens an alert and sends a notification to the aisle manager. Each record consists of the shelf, the time and how long the shelf stayed empty.
- Does out-of-stock detection need extra hardware?
- Not necessarily. The check runs on your existing store cameras, with no shelf sensors, smart shelves or electronic label hardware. Video can be processed on ALGI AI servers, or on an Algi Edge device on site if you want processing to stay on your premises. To bring a new aisle into scope, a camera that sees the shelf is enough.
- Are my existing IP cameras compatible?
- Standard IP cameras that provide an RTSP stream are compatible. What matters is how clearly the camera sees the shelf: cameras looking at the shelf face at a very steep angle or from too far away perform poorly. Share photos of your aisles and your camera layout, and we will work out which shelves are covered before installation.
- How long does installation take?
- A single store is set up in one day: camera streams are registered and shelf zones and thresholds are drawn. The first days run in observation mode to tune thresholds, so alerts are logged but no notifications are sent. Once the tuning has settled, notifications are switched on.
- Is out-of-stock detection compliant with GDPR and KVKK?
- The system looks at the shelf, not at shoppers. It does not perform face recognition, does not identify individuals and keeps no per-person records; the only thing measured is fill in a shelf zone. With an on-site Algi Edge device, live video stays in the store; with server processing, video is processed on ALGI AI servers. Your existing camera notice obligations remain unchanged.
Related rules from the library
ALL →DURATION
Checkout-Queue Detection
Instant alert to open another till when queue length passes the threshold.
CROWD_DETECTION
Heat-Map Analysis
In-store density map — which areas attract interest, which are dead zones.
OBJECT_DETECTION
Bin & Cleaning Triggers
Full bins and usage-based restroom cleaning notified automatically.
DON'T SEE YOUR NEED?
Dijital Atölye listens to the problem in the field, prototypes, validates and productizes. Some of the products running in the field today were born on this line.
Bring the problem, let's talk prototype →