We do not publish who we put the cameras in for.

The basic reason holds for any sector: we do not publish a client's name without their written authorisation. Not a phone call, not an «I don't suppose they'll mind»: written, from somebody with the authority to give it. That authorisation is asked for when it is needed, and until it arrives the name is not used.

But in security there is a stronger reason, and it is the one that makes this page what it is. Saying publicly who has hired us is saying who has a system built by us. With a bit of patience, quite a lot follows from that: what kind of installation they probably have, with which manufacturers, what it covers and what it does not. That is useful information for exactly the people who should not have it.

So the question worth asking when you see a security supplier's website full of logos is a simple one: if they show that client, what would stop them showing you? Sometimes the answer is that the client was happy to authorise it. And quite often the answer is that nobody asked.

This is not an excuse for having nothing to show. If you are evaluating suppliers and you want to talk to somebody who has us under contract, you can: ask, and we look for a case like yours —by type of installation, by size, by the problem underneath— and we ask that client whether they are happy to talk to you.

Two conditions, both obvious: the reference is given by whoever wants to give it, and it is given in the conversation, not on a web page. Whoever says yes says it knowing who they are going to talk to. It is slower than putting twenty-four logos in a grid and it is the only version that is not done at somebody else's expense.

By sector, the work is in private companies, banking, transport, factories and public administration, plus the venues where big events are put on, which is a trade of its own with different deadlines. It is not a list of ambitions: it is where the work we have done since 2006 is.

They have little in common on the outside and quite a lot on the inside. In all of them there is a space somebody has to control properly and not just on paper, with consequences if it fails: an entrance that must not open for the wrong person, an area nobody can go into at night, a count that has to be justified to somebody outside, a recording that has to exist the day it is asked for.

And in almost all of them there is a second thing: the system is not built from scratch on an empty plot. Something is already there, years old, from two or three different suppliers, half documented. Working with that —deciding what stays, what gets replaced and in what order— is half the trade.

The conversation works when there is somebody on the other side who has already had the problem land on them. Somebody who knows what a false alarm at three in the morning costs, because they have paid for it in patrols, in hours and in a night shift that stopped taking any notice of the system. Somebody who has had a camera pointing at a useless spot for two years and found out on the day it mattered. Somebody who knows why their insurer asks about the installer's certification.

With that person you can talk about the project in the first meeting, because they already know that what decides the result is not the make of the camera. And above all, they understand the word «no» when it turns up, which is what speeds a project up most.

The other half of the profile is that there is somebody inside responsible for the system, with a name: the head of security, of operations, of the plant, of maintenance. They do not have to be technical. They have to decide what is watched, who gets alerted and what changes when the work changes, and to tell us. The projects that are still useful five years on have that person; the ones that fade out did not.

First, and the one that comes up most: talking to somebody who understands. Not to a salesperson who takes notes and comes back a week later with what the office told them. In the first conversation you can argue about whether that camera is good enough to read a number plate, what happens with the backlight in that unit at seven, and why the recording days they were promised were not real. If something has to be checked, we say it has to be checked.

Second: a project before a quotation. What comes first is not a list of equipment with a total at the bottom, it is what is being protected, from what, what each point covers and what is left out. The quotation comes out of that, and that is why it can be argued about: when somebody wants to cut, you can see what is being taken away.

And third, the one least expected and most appreciated afterwards: being told no. That the camera is no good for what they want. That the detection is not solved today. That what they need is something else and it is not ours. That the deadline they are asking for is not real. Said in the first meeting it is a ten-minute nuisance; discovered at commissioning it is a quarter of a year.

The first step is not a sales meeting or a catalogue: it is an image from a camera you already have, or a visit if the site needs seeing. That is enough to say quite a lot: what can be seen from there and what cannot, what it would take to see it, and whether what you want to achieve is within reach of what is there or not.

That is where the first real answer comes from, and sometimes the answer is that it cannot be done. That is an answer too, and it arrives before anyone has spent anything. We think it is better to start like that than with a handsome proposal that falls apart at the first site visit.

If it can be done, it is done small before it is done big. A few cameras, one or two rules, running on your installation with your own people watching it, to check whether it detects what it is meant to and how often it gets it wrong. Nobody has to believe anything: you look. And it grows from there.

Maintenance comes with a calendar and with a report. Dates, the round, and what was checked point by point: which cameras were cleaned, which had lost their field of view, which alarms were genuinely tested, how many days of recording there are right now. A one-line annual report saying «installation inspected, correct» costs the same as an inspection and is not an inspection.

There is one check almost nobody does and that we put into the round: looking again at what can be seen from each camera. Installations move. A rack goes up that covers half the image, a tree grows, the traffic through the unit changes, somebody turns a camera to get a better look at the loading bay and takes out the area it used to cover. None of that raises an error: the system carries on recording, and recording something else.

And it grows in parts. Almost none of the big systems we maintain were built in one go: they started in one area, worked, and were extended the following year. It is slower to sell and it turns out better, because each extension is decided already knowing what worked in the last one.

With whoever compares three quotations by the total and picks the lowest. Not because looking at the price is wrong, but because those three quotations do not include the same things, and ours is going to be the expensive one: half of what we charge is the project done beforehand and the maintenance done afterwards, and that does not show up in the price per camera.

With whoever needs the system installed by a date that does not work, knowing that it does not work. And with whoever wants the paper more than the system: doing the bare minimum to get through an inspection. That can be done and somebody does it, but not us, because two years later the system is no use and the name in the documentation is ours.

When we see it coming, we say so in the first conversation. And if we know somebody who is what is actually needed, we say that too.