The cost of doing business
South Africa’s retail sector rang up an estimated R1.36 trillion in sales last year. Shrinkage – theft, fraud, and error – is estimated at around 2% of turnover. Do the maths and retailers are quietly writing off somewhere in the region of R27 billion a year, every year, as the cost of doing business.
Now here’s the part that should actually keep retail executives up at night: most of that loss was recorded, in full colour, on a camera the store already owned. Nobody watched it in time to stop it.
We built a surveillance industry to answer the wrong question
Retail CCTV was never designed to prevent anything. It was designed to prove something happened, after it happened, to whoever asks. That’s why 64% of retail managers only pull up footage once there’s already an incident to investigate. A complaint, a stock count that doesn’t add up, a till that’s short. The camera did its job. It just did the job of a filing cabinet, not a first responder.
Most stores run what the industry politely calls “exception-based surveillance”: footage rolls, nobody’s watching, and it only gets reviewed once something has already gone wrong. Most commercial CCTV is only kept for 30 to 90 days. Which means, in a huge number of cases, the evidence of exactly how and where a store is bleeding money is automatically deleted before anyone ever looks at it.
You spent the capital on the cameras, you’re paying for the storage and are, in effect, funding an archive of your own losses – and then erasing it on a rolling three-month cycle.
Break down the R27 billion and it gets worse
Internationally, roughly two-thirds of shrinkage splits between external theft and internal, employee-involved theft, with the remainder down to process and administrative failures, such as till errors, pricing mistakes and damaged stock never written off properly. Every single category on that list has a visible signature on camera, in the moment it happens: the item that goes into a bag instead of a basket, the stockroom door propped open after hours, the till drawer opened without a corresponding sale.
None of that requires a forensic review. It requires someone, or something, watching at the time. Almost nothing in a modern store is watching at the time.
And that’s before we even get to the queue
Loss prevention gets the security budget. Queue management barely gets a mention, which is strange, because it’s arguably costing retailers just as much. Up to 4% of total sales revenue disappears to queue abandonment. Nearly a third of shoppers will simply walk away from a purchase rather than stand in a long line, and more than three-quarters say they’re less likely to come back to a store after a bad checkout experience. Grow a queue from ten people to fifteen, and conversion at that till can drop by 10% – in real time, while a customer with a full basket decides your competitor’s line looks shorter.
That’s not shrinkage in the accounting sense. It’s revenue that never even made it onto the till slip to be stolen. And it’s sitting in the exact same video feed as the theft nobody’s watching.
Here’s the provocation: your cameras already know
Every one of these losses, from theft, to sweethearting at the till, to the queue that’s about to cost you a sale, is visible on footage you already have, on cameras you already paid for, as it’s happening. The infrastructure isn’t the gap. The question you’ve been asking that infrastructure is.
For twenty years, retail has asked its cameras: “what happened?” That’s a forensic question, and it only gets answered after the money is already gone. The question that actually protects revenue is: “what’s happening, right now, on camera fourteen?” That’s an operational question and it’s one your existing CCTV network can answer today, without a single new camera on the wall.
What this actually looks like on the floor
A queue-detection agent watching till lines doesn’t file a report at month-end. It flags the moment a line crosses your threshold and pulls a staff member onto an open register before the first customer walks out. A theft-detection agent doesn’t wait for a stock count to notice shrinkage. It catches the scan-avoidance or the bag-stuffing as it happens, while someone can still act on it. A traffic and heat-mapping agent doesn’t produce a quarterly footfall report nobody reads. It tells you that aisle three is empty and till two is backed up, so you can move a person before either one costs you a sale.
None of this replaces your floor staff or your loss prevention team. It replaces the assumption that a camera’s only job is to remember. Every one of these is the same footage, the same cameras, the same infrastructure you’ve already sunk cost into. Now, it’s just finally being asked a question that pays for itself instead of one that only helps you explain the loss.
So, why aren’t we doing this?
If the tool to stop a third of your shrinkage and a chunk of your queue-driven revenue loss is already mounted on your ceiling, the only real question left for South African retail is why it’s still being treated as a security expense instead of a live operations tool. The technology to flip that switch already exists and runs on the cameras you have. The only thing standing between the current R27 billion write-off and a materially smaller one is the decision to start asking your cameras a better question.
Author: Bruce Madgwick
Refraime’s Project Manager