The clock-in register at the gate is the single biggest source of payroll error in African private security. A guard signs the sheet at 06:00 — or somebody signs it for them. The sheet travels to HQ a week later. Hours are re-typed into a payroll spreadsheet. By the time the client queries an invoice, nobody can tell whether the guard was actually at the post.

GPS-based geofenced clock-in is the obvious fix. But "obvious" hides three years of operational learning. This is what actually works in African field conditions — across South Africa, Kenya, Nigeria, Uganda, Tanzania and Ghana — and what doesn't.

What a geofence really is

A geofence is a circle on a map. It has a centre (GPS coordinates of your post) and a radius (typically 50–200 metres). When a guard tries to clock in, the system reads the phone's GPS location and checks whether it falls inside the circle. Inside: clock-in accepted. Outside: rejected with the exact distance shown to the guard.

The radius is the parameter that matters most. Set it too tight (e.g. 20 m) and legitimate clock-ins fail in cloudy weather or dense buildings, because consumer-phone GPS drifts. Set it too loose (e.g. 500 m) and a guard at home down the road clocks in successfully. Most production deployments land at 100 m, configurable per site for the special cases.

The GPS reality on cheap Android phones

Most African private security guards use entry-level Android phones — Samsung Galaxy J, Tecno, Itel, Infinix. GPS accuracy on these in open conditions is 5–15 metres. In dense urban canyons (Sandton CBD, Nairobi CBD) it widens to 30–50 metres. Inside a parking structure or basement it can fail entirely.

This is why a 100-metre default works: it absorbs the drift in most conditions, while still proving the guard is at the post and not at home. For posts inside buildings — bank branches, embassies, shopping centres — supplement GPS with the building's Wi-Fi signature or an NFC tag at the desk.

Clock-in proof vs. patrol proof

A common mistake is treating clock-in as the only attendance signal. It proves the guard arrived at the post. It doesn't prove they stayed, didn't sleep, walked the patrol route, or were awake when the incident happened.

Patrol proof is a separate problem. The standard pattern is checkpoints — physical NFC tags or QR codes at fixed points around the site (gate, perimeter corner 1, perimeter corner 2, generator room, store room). The guard scans each one as they walk the round. The system records the scan time and the GPS at scan time, and flags rounds completed too fast or with checkpoints missed entirely.

If a guard tries to scan all checkpoints from the guard hut — common workaround — the GPS at scan time will be wrong, and the system flags it. Operations sees the pattern within a shift, not at the end of the month.

See the security operations page for the full patrol-management capabilities, including SOS panic button and digital OB.

Anti-buddy-punching: what actually stops it

Buddy-punching — one guard clocking in another — is the oldest attendance fraud in the world. Three controls reduce it almost to zero:

SIM-tied clock-in: the clock-in is bound to the SIM card registered to the guard. A different phone with a different SIM cannot clock them in.
Optional selfie at clock-in: the camera captures the guard's face at the moment of clock-in. Operations spot-checks; no facial recognition is needed for this to work as a deterrent.
GPS at clock-in: the location proves the device was at the post. A guard at home can't clock in for a colleague who's at the post — the locations don't match.

For sites with biometric devices already installed (ZKTeco, Suprema, Anviz, Hikvision), the biometric reading is the strongest control. The GPS app and the biometric device can coexist — many security companies run both, depending on the post.

USSD for guards without smartphones

Not every guard has a smartphone or a data plan. USSD attendance solves this: the guard dials a short code (e.g. *123#), picks "clock in" from the menu, and gets an SMS confirmation. The SIM is tied to the registered guard. No app installation, no data plan, no smartphone required.

USSD doesn't carry GPS — that's its limitation. So a USSD clock-in is taken on trust that the SIM is at the post. For low-risk posts (overnight relief at a residential gate) that's acceptable. For high-risk posts (cash-in-transit holding points, large industrial sites) you should require GPS or biometric.

What regulators care about

In South Africa, PSIRA (the Private Security Industry Regulatory Authority) doesn't mandate a specific attendance system. What they audit is whether the guard you billed for at a site was actually deployed there, whether their PSIRA registration was current, and whether your payroll matches the deployment record. A geofenced clock-in log with site, time, and guard ID is the cleanest defence in an inspection.

In Kenya, the Private Security Regulatory Authority (PSRA) is similar — they want to see deployment records and matching payroll. The CID clearance and Ministry of Interior licensing add a documentation layer. Both are easier to maintain in a system with expiry alerts than on paper.

In Nigeria, NSCDC and state Ministries of Internal Affairs license private security companies; guard-by-guard CID clearance is the documentation layer. The same pattern holds: deployment records, payroll consistency, current licences.

The month-end change that actually saves time

The single biggest operational change is the speed of month-end. With paper registers:

Day 1–3: collect register photos from 80 sites via WhatsApp.
Day 4–6: re-key into a spreadsheet, reconcile to the schedule.
Day 7: compute OT, allowances, deductions.
Day 8: build per-client invoices.
Day 9: run payroll. Fix errors.
Day 10: payslips issued.

With geofenced clock-in feeding payroll directly:

Day 1: attendance is already in the system because clock-in happened at the post.
Day 1: OT computed automatically against shift template.
Day 2: per-client invoices generated from same data.
Day 2: payroll review and approval. Payslips live.

That's the change. Three hours instead of three days.

What to ask before signing

Is the geofence radius configurable per site, or one global value?
What happens when GPS fails or a phone breaks — can operations override with an audit trail?
Does the patrol checkpoint scan record GPS at scan time, or only the scan itself?
Is USSD attendance supported for guards without smartphones?
How does the system handle multiple shifts per day, and night-shift roll-overs across the date boundary?
Does the same attendance data drive both payroll and per-client invoicing?

If you want to see this run on your sites, your shifts and your client contracts, book a working demo — bring 2-3 active sites and we'll set up live geofences for them inside the session. Or jump to the security operations page for the full feature breakdown.

Geofenced Clock-In for African Security Companies: What Actually Works in the Field