Most pilferage on an industrial site is not dramatic. It is small, repeated and procedural — a routine that works because nobody is required to verify anything at the moment it matters. That is why adding cameras alone rarely moves the number.
What reduces losses is closing the specific gap where material can leave without a record. Cameras make that gap visible and enforceable; they do not close it by themselves.
Pilferage is a process problem first
Ask a simple question about your site: at which points can material physically leave, and at each of those points, what record is created, by whom, and who checks it against something independent?
Wherever the answer is “a register kept by the same person who controls the gate”, you have found the gap. It is not an accusation; it is a design fault. A control that depends on one unsupervised person is not a control.
The four places losses actually happen
- The gate, during vehicle movement. The largest single exposure on most sites. Loaded vehicles leaving against paperwork nobody reconciles, and returns that were never actually returned.
- Stores issue and return. Material issued against a job that consumed less than was drawn, with no return step that anybody enforces.
- Scrap and waste movement. Consistently the most under-controlled route on an industrial site, because scrap is treated as worthless right up to the point where it is sold. Valuable material leaves as scrap.
- Contractor and shift-change movement. Periods of high foot traffic where the gate is busy, verification is rushed, and personal bags move in volume.
Notice that three of the four are gate or stores events, not shop-floor events. Systems that concentrate cameras on production and thin them at the gate are covering the wrong ground.
What cameras can and cannot do here
Cameras do three things well in this context, and one thing badly.
- They make a record that is independent of the person keeping the register. This is the real value, and it changes behaviour on its own.
- They let a dispute be resolved with evidence rather than with an argument between two departments.
- They deter, where they are visibly positioned at the point of decision — the gate, the weighbridge, the stores counter.
- They do not detect pilferage. Nobody watches live footage, and nobody reviews thirty days of it speculatively. Footage answers a question you already have a reason to ask.
Which means the useful design question is not “where do we put cameras” but “which record will we reconcile, and which camera makes reconciliation possible when the numbers disagree?”
The controls that reduce losses fastest
- Camera at the gate covering both the vehicle and the paperwork exchange, positioned so the number plate is readable and the transaction is visible in the same frame or on a second camera at the same point.
- Independent gate record. Access control at the gate produces a timestamped log nobody at the gate can edit, which is the point.
- Bring the weighbridge into scope. A weight record plus a plate image plus a gate log is a reconciliation that is very hard to defeat.
- Cover scrap movement explicitly, with the same rigour as finished goods. This one change frequently finds the loss.
- Biometric attendance for contractors, so headcount at the gate matches the people actually on site — also the fastest way to remove buddy-punching from payroll.
- Retention that matches your audit cycle. If discrepancies surface at a monthly or quarterly stock-take, thirty days of footage may already be gone. This is the most commonly missed requirement on an industrial site.
How to know whether it worked
Decide the measure before you install, not after. Usually it is the variance at stock-take, or the value of the gap between issued and consumed material for a defined line, measured for a few cycles before the change and the same number after.
Without that baseline you will have opinions. With it you will know, and you will also know which of the changes did the work — which matters, because the next site or the next line can then get the effective control first.
We would rather be judged on that number than on the camera count, and it is a reasonable thing to ask any vendor to agree to before they quote.
Common questions
Will cameras alone reduce pilferage?
Partly, through deterrence at visible points like the gate and stores counter. But footage only helps once somebody has a reason to look, so the reduction comes from pairing cameras with a record that gets reconciled - a gate log, a weight, an issue-and-return step. Cameras make the control enforceable rather than being the control.
Do we need to tell staff the cameras are there?
Yes, and it works in your favour. Visible cameras and a clearly communicated policy deter far more than covert ones, and they avoid the industrial-relations problem that discovering hidden surveillance creates. Deterrence is the return you are buying; secrecy reduces it.
How long should a factory keep footage?
Long enough to cover your audit cycle plus the time it takes to investigate. If stock-take is monthly, thirty days is already marginal because the discrepancy is found at the end of the period. Many plants are better served by sixty days, which is a disk sizing decision made at the design stage.