Guard attendance in South Africa: why we are building it on WhatsApp, not an app
Ask a security company owner what they need from software and you will get a long list: patrol tracking, incident reports, PSIRA compliance, occurrence books, client reporting.
Ask a site manager at ten past six on a Monday morning and the list collapses into one question. Did the guard arrive?
We spent time inside the operations of a KwaZulu-Natal security company running multiple sites, and that answer kept coming back. Not analytics. Not dashboards. Whether the person who is supposed to be at the gate is at the gate, and if not, how quickly somebody knows about it.
Everything below is what that conversation did to our product plan.
The industry sells the wrong thing first
Sit through any guard-management demo and you will be shown a mobile app, usually paired with a rugged handset for each post, sometimes a biometric clock bolted to a wall. It works. It is also the reason most small and mid-sized security companies are still running attendance out of a spreadsheet.
Count the real cost of the device approach across a twelve-site company:
- a handset per post, plus a monthly fee on each one
- chargers and cables, including the ones that walk away
- data for every device
- a breakage budget, because these things live outdoors, overnight, in the rain
- training every new guard on an app they will touch twice a shift
- and the failure nobody demonstrates: when the device is flat, lost or locked in the office, the guard cannot clock in at all. The shift goes unrecorded and the controller is back on the phone.
That is before the software licence. For a company billing on thin margins per guard hour, the arithmetic usually stops right there.
Meanwhile, the real system is already running
Here is what those companies actually do today, almost without exception.
The roster lives in Excel. A manager or a controller builds the monthly schedule there, because they know Excel, they are already paying for it with the Microsoft subscription, and it has outlasted every software salesman who ever walked through the door.
And the guards are told where to be over WhatsApp. A group, a direct message, or a controller working down a list by phone.
So the channel is already chosen, already installed on every guard's phone, already understood, and it costs nothing to roll out. It simply is not recording anything. The information evaporates the moment it has been read.
That is the gap. Not that these companies need dragging into the digital age. They are already there. Nothing is being written down.
Why WhatsApp wins for attendance specifically
Five reasons, in the order an owner cares about them.
- Nothing to install. Every guard already has it. No app store, no eighty megabytes off their own data, no setup queue on day one of a new contract.
- No device to buy. The phone is already in the guard's pocket, already charged, and already theirs to look after.
- No training. If someone can send a message, they can go on duty. That is the whole learning curve.
- It runs on what they actually carry. Entry-level Android, older handsets, tight data budgets. WhatsApp was built for exactly that, which is why it won South Africa in the first place.
- It is where the shift conversation already happens. Checking in does not become a new habit. It becomes a reply.
What a check-in actually looks like
Nothing clever. Clever is what breaks at three in the morning.
Fifteen minutes before the shift starts, the guard gets a message naming the site and the shift. They tap one button: on duty. The system stamps the time, the site and the guard's identity, and writes it to the log. Going off duty works the same way at the end.
If no check-in arrives by the end of a grace period, the controller gets an exception message: which guard, which site, which shift, how late. That is the entire product for version one. Not a dashboard nobody opens. One message, when something is wrong.
Two honest limitations, because they will come up in the first sales meeting.
Location. WhatsApp cannot track anyone in the background, and it should not be able to. A guard can share their location along with a check-in, but that is a deliberate act and it is a single pin at a single moment, not continuous tracking. If a client contract requires proof of patrol routes through the night, WhatsApp check-in is not that, and it is still a job for patrol points with QR or NFC tags. Attendance and patrol proof are two different problems. Conflating them is how companies end up buying far more than they need.
Identity. A phone can be handed to somebody else. So can a card be shared and a fingerprint reader be fooled. Every attendance system sits somewhere on that spectrum. Pairing each guard to a registered number, and adding a photo or a location pin on check-in where a client demands it, gets most operations where they need to be without doubling the cost.
The part that actually decides adoption: you keep your spreadsheet
This is the piece we had wrong at first, and it matters more than all the design above.
Every previous attempt at selling these companies software asked them to abandon a roster they have been refining for years and re-enter all of it inside somebody else's interface. That is not a software decision, it is an operational risk, and the answer is almost always no. Or worse, a polite yes followed by nothing changing.
So the brief became: do not replace the spreadsheet, read it. The roster stays where the controller built it. The system takes the shifts from it, sends the check-ins, records what comes back, and writes attendance into a sheet they open exactly the way they always have. The manager's Monday does not change. The information just stops disappearing.
Software that demands a new habit gets abandoned in week three. Software that runs alongside the habit already there gets used.
What it costs to run
Two costs, and they behave differently.
The build is a project cost. The running cost is Meta's rather than ours: WhatsApp is billed per message by category, and a business-initiated message such as a shift notification is priced differently from a reply inside a conversation the guard started. Rates change and they vary by country, so check Meta's current pricing rather than trusting any blog post, this one included.
One design consequence is worth knowing up front. The cheapest attendance system is the one the guard initiates. A guard messaging in to go on duty opens a conversation window that costs materially less than the company pushing a template out to every guard for every shift. Build it the wrong way round and at two hundred guards it is the message bill, not the software, that kills the project.
Where this goes beyond security
Guards are the sharpest version of a problem that turns up everywhere: staff who start work somewhere other than your office.
Cleaning teams arriving at client premises. Plumbers and electricians on call-outs. Delivery drivers. Merchandisers in retail. Contract catering. Home-care nurses. In every one of them the manager's first question of the day is identical — did they arrive, and if not, who needs to know right now — and in most of them the answer still comes from phoning around.
The mechanics transfer almost unchanged. The roster differs. The question does not.
Where we are with it
To be straight about status: this is the design we are building now, not a finished case study. It came out of the operations of a real security company and was tested against how their controllers and managers actually work. It replaced an earlier plan of ours that involved an app and a device at every post. Once it has run through a full pilot we will publish what actually happened, including the parts that did not work.
If you run shift workers and your attendance still depends on somebody phoning around at six in the morning, message us. The conversation is free, and you will be talking to the kind of assistant we build.