Video surveillance system options

Every location has its own security challenges, and the video surveillance system that fits one site may not fit another. The choice matters because the wrong model either ties up your money in hardware or leaves you without anyone watching the feeds.

This guide compares the main ways to run video surveillance — fully hosted, hybrid, and remote-monitored — and what to check when you choose.

The three deployment models

Fully hosted (video surveillance as a service)

A vendor manages the whole system for you: cameras, analytics software, cloud storage, and operations under a service agreement, with remote monitoring of your premises included.

• Benefits — simple to set up and run with limited technical effort on your side; no upfront hardware cost (you pay a monthly fee based on usage); scales easily as you add or remove cameras; footage accessible remotely from any device.
• Trade-offs — you own none of the hardware, have limited control over it, and if the contract ends you lose access to the infrastructure built by the vendor. Monthly costs continue for as long as you use it.

Best fit: organisations that want a working system without running a security infrastructure team, and that value uptime over hardware ownership.

Hybrid hosted

You buy the cameras and recorders on site, and connect your hardware to a cloud server where analytics and storage run. You can monitor the locations with your own team or with an agency.

• Benefits — lower monthly fees (you own some of the hardware and pay mainly for cloud and monitoring); more control over hardware, and you are not forced into a vendor’s limited camera range.
• Trade-offs — upfront hardware cost; the hardware set changes as you add cameras; you must run a small team to keep the hardware working.

Best fit when you already have an in-house team that can maintain cameras, and you want to keep the capital investment and the control.

Remote-monitored surveillance

Security professionals watch your live footage from an off-site location — they monitor for suspicious activity, log daily events, and connect with on-site personnel when needed.

• Benefits — round-the-clock monitoring at lower cost than staffing a guard post on every site; monitoring can be scoped to specific locations and hours to control expense; trained operators respond quickly with the help of local staff; advanced analytics can be layered on the feeds.
• Trade-offs — the location needs a reliable internet connection for live feeds, which can be a problem in remote areas.

Best fit when the value is in having someone actually watching the feeds, not just recording them — especially across multiple sites.

Comparing the three models

The models differ less in what the cameras do than in who owns the equipment, who watches, and what happens when something fails.

	Fully hosted	Hybrid	Remote monitored
Upfront hardware cost	Low	High	Depends on existing estate
Monthly cost	Higher, ongoing	Moderate	Scales with sites and hours
Who maintains the kit	Vendor	You	Vendor, or you plus vendor
Hardware ownership	Vendor	You	Either
Who watches	Vendor’s monitoring	You or your agency	Off-site monitoring team
Switching cost	Higher — you leave the infrastructure	Lower — you own it	Lower
Weak point	Vendor dependence	Maintenance burden	Connectivity at the site

A useful way to choose: if the problem is that nobody watches, remote monitoring addresses it directly. If the problem is no system at all, hosted gets you running fastest. If you already have cameras and in-house capability, hybrid protects the investment you have made.

The fundamentals that matter whichever model you pick

Model choice is secondary to three decisions that determine whether the system is worth anything:

• Coverage matched to the risk. Cameras at the gates, doors, high-value stores and loading areas — the points where something entering or leaving can be seen and identified. Wide shots of empty aisles record activity without identifying anyone in it.
• Retention long enough to be useful. Footage that expires before you notice the loss is footage that never existed. Agree the retention period against how quickly you would realistically detect a problem, not against the cheapest storage option.
• Image quality where it counts. The purpose of most cameras is to identify a person or a vehicle after an event. If the footage cannot do that at the distance and lighting it operates in, the installation looks complete and proves nothing.

Footage is personal data

Recorded images identify real people, so footage is personal data and brings obligations alongside the security benefit. Three habits keep the system defensible:

• Decide a retention period and honour it. Keep footage for as long as a genuine security or evidential need requires, and delete it after that rather than accumulating it indefinitely.
• Restrict who can view what. Access should follow the same logic as physical access — people see the footage they need for their role, and views are logged so the system itself is accountable.
• Tell people they are recorded. Staff, and visitors where relevant, should know that cameras are operating and for what purpose. Signage is the ordinary way to do this, and it removes most of the objection.

Apply the same discipline to exports and clips: a downloaded incident video leaves your controlled system, so decide where clips go and who authorises sharing them. Where footage leaves the premises to a monitoring provider or a cloud service, the contractual arrangements on storage, access and deletion belong in the agreement rather than being assumed.

What to check when you choose

Whichever model you pick, the system is only worth what it does when something happens:

• Reliability — who keeps the cameras online and recording? Failed cameras and missing maintenance contracts are the classic cause of dead footage when you need it most.
• Monitoring — is anyone actually watching alerts around the clock, or is the system recording with nobody responding?
• Flexibility — can the system scale up and down without ripping out hardware, and do alerts stay fast as it grows?
• Incident response — are there standard operating procedures for what to do when an alert fires: when to inform you, when to contact authorities, and what the escalation steps are?

The right system is the one that produces footage and responses you can act on, matched to your budget. For the decision on whether a managed provider is right for you, see how to choose a security agency and what remote security monitoring involves.

Frequently asked questions

Do we need someone to watch CCTV, or is recording enough?

Recording alone is evidence after the fact. It does not stop a theft in progress, challenge a person at a gate, or let anyone respond while something is still happening. If the risk you are managing is loss rather than investigation, someone needs to see the alert and act on it — on site, or through a monitoring team.

How long should CCTV footage be retained?

Long enough that you would still have it when the loss is discovered. That depends on your operation: a warehouse reconciling stock monthly needs a different retention period from a retail store checking daily. Agree it against your detection cycle rather than the cheapest storage tier.

Is cloud storage less secure than recording on site?

Not inherently, and on-site recording has its own weakness — a thief who reaches the recorder takes the evidence with it. What matters is access control, encryption and where the data physically resides. Ask each vendor where footage is stored, for how long, who can access it, and what happens to it if you end the contract.

What happens to the system if we change provider?

Under a fully hosted model, very little of it is yours: the cameras, storage and analytics sit with the vendor, so ending the relationship means starting again. Under a hybrid model you keep the hardware. If switching cost matters to you, that single question should shape the choice more than the monthly fee.

How many cameras do we need?

Start from the risk points rather than a ratio. Map the entrances, gates, high-value stores and loading areas, decide what you need to see at each, and size the installation to that. A coverage plan is also what lets you demonstrate, later, whether an area was genuinely covered.

Does video analytics replace human monitoring?

It changes the job rather than removing it. Analytics raises an alert when something crosses a line or enters an area at an unusual time, which lets a smaller team cover more ground than watching many live feeds at once. The alert still needs a human to verify and respond — the video analytics guide covers how that works in practice.