Centralised CCTV across a branch network: lessons from the field

A bank or insurer with sites across a country doesn’t have a camera problem. It has a coordination problem. Each branch can have perfectly good CCTV and the organisation as a whole can still be blind — because no one can see across the estate, footage can’t be retrieved when it’s needed, and an incident at one site teaches the others nothing.

Centralising surveillance across a branch network is one of the most common engagements we take on. Here’s what we’ve learned about doing it well — and the places it tends to go wrong.

The goal is one operating picture, not more cameras

The temptation is to measure success in hardware: number of cameras, resolution, hours of storage. Those matter, but they’re inputs. The output that matters is a single operating picture — the ability to see the security state of the whole network from one place, and to act on it.

That reframing changes the design. Integration — tying CCTV together with access control, fire detection and evacuation systems under one operating view — usually delivers more value than adding cameras to any one site. An operator who can see that an access-control alarm and a camera event happened at the same door at the same time has context that ten more cameras wouldn’t provide.

Bandwidth is the constraint everyone underestimates

Branch networks rarely sit on fat, reliable links — especially across regions. A design that assumes every camera streams full-resolution video to a central point in real time will choke the network and frustrate everyone.

The workable pattern is almost always record locally, review centrally: full-resolution footage stored at the edge, with the central site pulling lower-bandwidth streams for live monitoring and full-quality clips on demand. It’s less elegant on a diagram and far more reliable in practice. Design for the network you have, not the one in the brochure.

Retention and retrievability are a policy decision, not a default

“How long do we keep footage?” is usually answered by whatever the recorder shipped with. It should be answered by your risk and your obligations. Fraud investigations in financial environments can surface weeks after the event; a 7-day default is worthless when the dispute arrives on day 20.

Two questions decide the storage design: how far back might you genuinely need to look, and how quickly must you produce footage when asked? Get those answered before sizing anything.

Standardise the estate, tailor the site

Across many branches, consistency is a feature: the same camera families, the same naming, the same recording policy, so an operator moving between sites isn’t relearning the system each time. But — and this is where a baseline assessment earns its place — the coverage at each site should follow that site’s risk. A high-cash, high-footfall branch next to a known risk is not the same problem as a quiet records office. Standard platform, site-specific coverage.

The unglamorous things that decide success

Once the design is right, projects tend to live or die on operational details that never make the proposal:

• Time synchronisation. If cameras, access control and alarms don’t share a clock, correlating an incident across systems becomes guesswork. Sync everything.
• Maintenance cadence. A camera that’s been dark for three weeks is discovered at the worst possible moment. Build monitoring of the monitoring.
• Who’s actually watching. A centralised system with no one trained to use it is a very expensive recorder. The people and the post orders matter as much as the platform.
• Commissioning, properly. “It’s installed” is not “it works.” Each site should be verified against the design before it’s signed off.

What it looks like done right

At its best, centralisation takes an estate of disconnected, individually-fine sites and turns it into something an organisation can actually run: branches, ATMs and entry points unified under one monitoring picture, integrated with access and fire, with footage that’s there when it’s needed. The cameras were never the hard part. The system around them is.

Planning surveillance across multiple sites? Request an assessment — we design the operating picture first, then the kit that serves it.