8 April 2026 · 6 min read
How many CCTV cameras does your site actually need?
Fewer than most quotations assume. Count the decisions you need footage to settle, and each one becomes a camera position. Everything else is a camera watching a wall.
Count decisions, not square feet
Start from the arguments you need footage to settle. Who came through the gate at 9.40pm. Who was at the till when the drawer came up short. What happened at the weighbridge before the load was signed for. Each of those is one camera position, and the list is usually shorter than a vendor's quotation.
Quotations built from floor area instead produce cameras watching empty corridors at full resolution, filling disks with footage nobody will ever request. That is where the money goes on an oversized system: not in the cameras, but in the recorder and disks needed to store what they see.
Retention is a number you choose, not a limit you accept
How many days you keep comes down to three things you control: how many channels, at what resolution, onto how much disk. A recorder shipped with a default fifteen days is not a constraint, it is a configuration.
Set retention against your own discovery cycle. Retail shrinkage usually surfaces at month-end stock count, so seven days of footage is useless and thirty to sixty is the honest requirement. A school investigating a corridor incident normally hears about it within a week. Ask for the disk capacity for 15, 30 and 60 days and choose deliberately.
The lens matters more than the megapixels
A high-resolution camera pointed with the wrong lens at the wrong light produces a grey shape where a face should be. The gate at night, a corridor lit from one end, a shop floor under high-bay lighting and a forecourt in direct afternoon sun are four different optical problems.
This is the single most common reason a system disappoints after handover. It is decided at survey, on site, in the light the camera will actually live in — not from a floor plan in an office.
Insist on being able to find the footage
A system nobody can search is a system that does not work. When an incident happens, the person on duty needs to find a time and a channel, export a clip and hand it over, without calling the installer.
Ask for a channel map and thirty minutes of training at handover. If the person who will actually use it cannot export a clip unaided at the end of that, the handover is not finished.
- A floor plan with each camera numbered and its field of view marked
- Admin credentials in writing, held by the organisation rather than one person
- The retention period the system was actually configured to achieve
- Named staff shown how to search, export and hand over footage
Also asked
Follow-up questions
Is a higher-megapixel camera always better?
- No — beyond a point it costs you retention rather than buying clarity. Every extra megapixel fills disk faster, so an oversized resolution shortens how many days you keep. Match resolution to the identification distance you need, then spend what is left on disk.
Should cameras cover the inside of guest rooms or washrooms?
- Never. Those areas are excluded on privacy grounds regardless of what a system is technically capable of. Coverage belongs at entrances, corridors, parking, tills, stores and receiving points.
Can we add cameras to a system we already have?
- Usually, if the existing recorder has spare channels, spare disk and speaks ONVIF. That is worth checking before replacing anything — an expansion is frequently a switch and a disk rather than a new system.