Most businesses buy cameras the way they buy a fire extinguisher: once, in a hurry, and then never think about them again until the day something goes wrong. Then comes the discovery that the footage is too grainy to identify anyone, the angle missed the actual incident, or the recordings only went back three days and the theft happened last week. The cameras were on the wall the whole time. They just were not doing the job everyone assumed they were.

A camera on the wall is not a security system. It is a camera on the wall. The system is everything around it, where it points, what it records, how long it keeps it, who is alerted, and what happens next.

The good news is that camera technology has quietly leapt forward. The boxes look much the same, but what sits behind them, sharper sensors, real network video, and on-device analytics, has changed what a camera system can do for an ordinary business. The trick is knowing what is genuinely useful and what is sales-sheet theatre.

What actually changed

Old analogue CCTV did one thing: it pointed a fuzzy camera at a spot and dumped the picture onto a recorder in a back room. If you were lucky, the tape had not been recorded over. Modern systems are IP-based, which simply means each camera is a small networked device. That one change is what makes everything else possible, because a networked camera can be viewed from anywhere, can record to resilient storage, and can run software that understands what it is looking at.

That last part is the real shift. A modern camera does not just see motion, it can tell the difference between a person, a vehicle, and a cat walking past. That sounds like a gimmick until you have spent a weekend deleting four hundred alerts triggered by moths and shadows.

What modern CCTV can actually do

Strip away the jargon and here is where today's systems genuinely earn their place:

A camera you have to scroll through after the fact is a record. A camera that tells you something is wrong while it is happening is a system. The gap between those two is the whole point of upgrading.

The questions that actually decide quality

Resolution gets all the attention on the brochure, but the things that decide whether a system is any good are more boring and more important. Before you buy anything, work through these:

  1. Coverage and placement. Ten cameras pointed at the wrong things are worse than four pointed at the right ones. Map your genuine risk points first, entrances, tills, stockrooms, gates, then place to cover them with no blind spots.
  2. Low light and glare. Most incidents happen at night or against a bright doorway. A camera that produces a beautiful image at noon and a useless smear at 7pm has failed at its actual job. Ask specifically about low-light and backlight performance.
  3. Storage and retention. How many days do you keep, and what happens when the disk fills? Decide your retention period deliberately, then size the storage for it, do not let the recorder quietly decide for you.
  4. Network and bandwidth. Cameras are now part of your IT, not a separate world. They need to sit on a properly considered network so they do not choke your connection or, worse, become an open door into it.
  5. Who actually watches it. Footage nobody reviews is just an expensive hard drive. Decide upfront whether someone monitors live, whether alerts go to a phone, and who acts when one fires.

Start with your single worst blind spot

You do not have to wire the whole property at once. Find the one place where an incident would hurt most and where you currently have no clear view, the cash office, the loading bay, the main gate, and get that one right. A single well-placed, well-recorded, well-monitored camera beats a dozen that nobody can rely on.

Cameras are data, so POPIA applies

It is easy to forget that CCTV footage of identifiable people is personal information, which means POPIA has an opinion about it. You do not need to be nervous, you need to be deliberate. Put up visible signage that recording is taking place, be clear about why you are filming, keep footage only as long as you genuinely need it, and control who can access it. Pointing a camera over the fence onto a neighbour's property or the public street is exactly the sort of thing that turns a security measure into a complaint. Treating footage as the sensitive data it is keeps you on the right side of the line.

Where it ties into everything else

This is the part most camera installers miss, because they sell boxes, not systems. Cameras are one layer of a security picture that also includes access control, alarms, your network, and the servers that store it all. When those pieces are designed together, a forced door can summon the nearest camera, an alarm can be verified before a response is dispatched, and footage is backed up somewhere it survives the very break-in it recorded. When they are bought piecemeal from whoever was cheapest that month, you get a wall of cameras that each work alone and a business that is still, somehow, not secure.

That is the difference between buying cameras and designing security. The hardware is the easy part. Knowing what to watch, how it should fit together, and what should happen when something trips, that is the part that actually protects you.