Quick Answer
Geofence attendance is a clock-in method that uses a virtual boundary drawn around a site’s GPS coordinates. A guard can only mark attendance when their phone confirms they are physically inside that boundary, so the record shows not just when someone started a shift but where they were standing when they did.
What Geofence Attendance Actually Is
A geofence is a circle — occasionally a polygon — drawn on a map around a fixed point. In security operations that point is a site: a gate, a lobby, a plant entrance, a residential tower. The geofence is the area within an agreed distance of it.
Geofence attendance ties the act of marking attendance to that boundary. When a guard opens the app and taps to start their shift, the phone reports its coordinates. If those coordinates fall inside the fence, the check-in is accepted and stored with a timestamp and a location. If they fall outside it, the check-in is refused, or accepted and flagged, depending on how the system is configured.
That is the whole mechanism. It is not complicated, and its simplicity is the point. The difference between a paper muster roll and a geofenced check-in is not sophistication — it is that one of them is a claim and the other is a record.
Why “attendance” and “patrol” are not the same thing
This distinction matters and gets blurred constantly, including by vendors.
Patrol verification answers: did the guard walk the rounds? It works through checkpoint scans distributed across a site — NFC tags, QR codes, GPS waypoints — each scan producing a timestamped record at a specific location. It is about coverage over the course of a shift.
Attendance answers a narrower question: did the guard turn up, when, and did they stay for the contracted hours? It is about presence at the boundaries of the shift.
You need both, and they produce different evidence. A guard can pass every checkpoint and still have arrived ninety minutes late. A guard can clock in perfectly on time and never leave the guardhouse. Systems that treat attendance as a by-product of patrol data tend to under-serve payroll and over-serve operations. If you are being billed or billing on hours, attendance needs to stand on its own.
For the patrol side of this, our guide to the guard tour system covers checkpoint verification in detail, and the security guard tracking system piece covers real-time alerting on missed rounds.
The Problem It Was Built to Solve
Ask anyone who has run a multi-site guarding contract what their attendance disputes look like and you will hear a version of the same four stories.
Buddy punching
One guard marks attendance for another. On a paper register this requires nothing more than a second signature and a cooperative colleague. On a shared biometric device at a single gate it is harder but not impossible. With a personal device inside a geofence it becomes materially more difficult, because the phone has to be physically at the site — though, as covered below, it does not become impossible, and any vendor telling you otherwise is overselling.
The retrospective muster roll
The register gets filled in at the end of the week from memory. Every entry is neat, every time is a round number, and none of it is evidence of anything. This is the most common failure and the least dramatic — nobody is committing fraud, they are catching up on admin. It still produces records that collapse the moment a client questions them.
The late start nobody records
A guard due at 06:00 arrives at 06:40. The supervisor knows, the guard knows, and the register says 06:00 because writing anything else creates a conversation neither of them wants. Over a month across forty guards this is a substantial and completely invisible number of unworked hours — hours the client is being invoiced for.
The disputed exit
A guard leaves early. Two weeks later the client asks why the site was unmanned at 21:15 on a Thursday. Without a check-out record there is no way to establish what happened, and the contractor absorbs the doubt.
None of these are exotic. They are the ordinary friction of running a labour operation on paper, and they all share a root cause: the attendance record is produced by the same person whose attendance is in question, with nothing independent attached to it.
How a Geofence Attendance System Works, Step by Step
The flow is short. Complexity here is a warning sign, not a feature.
1. The site is mapped. An administrator sets the site’s coordinates and a radius. This is usually done once during onboarding and revisited when a site’s layout changes.
2. The guard is assigned. Attendance is only meaningful against an expectation — a roster entry saying this guard, this site, this shift. Without a roster the system records that someone was somewhere, which is not the same as knowing whether the site was covered.
3. Check-in. The guard opens the app at the start of the shift and marks attendance. The device supplies coordinates; the system compares them to the fence.
4. The record is written. Accepted check-ins store the time, the coordinates, the guard, the site, and the shift. This record is not editable by the guard afterwards — that immutability is the entire value.
5. Exceptions surface. Out-of-fence attempts, late check-ins against roster, and missing check-ins become supervisor alerts rather than month-end discoveries.
6. Check-out closes the shift. The same process in reverse. Without an enforced check-out, hours totals are estimates, and the on-site headcount is wrong.
7. The data becomes a report. Timesheets, exception summaries and client-facing attendance reports are generated from the stored records rather than reconstructed by hand.
The step most operations get wrong is step 5. Collecting the data is easy; nothing improves unless somebody looks at the exceptions in the same week they occur.
Geofence vs Biometric vs Paper: What Each One Actually Proves
Each method proves something different, and the honest comparison is about evidence, not features.
| Paper register |
No |
No |
No |
Near zero |
Every entry is an unverified claim; trivially backdated |
| Fixed biometric device |
Yes |
Yes, but only at the device |
Yes |
Hardware per site; install and maintenance |
Fixed to one point; fails on multi-entry or roving sites; queues at shift change |
| Geofence attendance (app) |
Weakly — proves the device was present |
Yes, to within GPS accuracy |
Yes |
Uses the guard’s existing phone |
Depends on GPS quality; device can be handed over |
| Geofence + photo/selfie capture |
Yes, in practice |
Yes |
Yes |
Same as above |
Adds friction; storage and privacy obligations |
The row worth sitting with is the third one. A geofence proves a device was inside a boundary. It does not, by itself, prove which human was holding it. If someone hands their phone to a colleague, the geofence is satisfied. This is why systems that care about identity pair location with a capture step — a selfie at check-in, or a biometric on the device itself.
Vendors rarely volunteer this. It is worth asking directly, because the answer determines whether your attendance data survives a serious dispute or merely a casual one.
Where each one belongs
- Single fixed entrance, high headcount, one shift change point — a fixed biometric device is genuinely good and probably cheaper over time.
- Multiple sites, roving guards, sites without power or mounting points — geofence attendance, because hardware per site does not scale and roving guards have no device to walk to.
- High-dispute or high-liability contracts — geofence plus photo capture, because you will eventually need to prove identity, not just presence.
- Very small single-site operations where everyone knows everyone — paper is defensible right up until the first contract that requires evidence, at which point it isn’t.
Setting the Radius: The Decision Everyone Gets Wrong First
The radius is the single configuration choice that determines whether the system helps or becomes a daily irritation.
Too tight, and legitimate check-ins fail. GPS on a consumer phone is not surveyor-grade; accuracy varies with sky visibility, building density, weather and handset age. A fence of twenty metres around a gate in a dense urban block will reject guards standing at that gate, repeatedly, and the guards will conclude the system is broken. They will be more or less right.
Too loose, and the fence stops meaning anything. A five-hundred-metre radius around a city-centre site includes a café, a bus stop and someone’s flat. The check-in is technically location-verified and practically worthless.
Sensible practice:
- Start wider than feels right, then tighten with observed data rather than guesses. A fence that rejects nobody in week one tells you the ceiling; you can lower it from there.
- Size to the site, not to a policy. A ten-acre industrial park and a lobby desk do not get the same number.
- Look at the rejection log before blaming guards. Repeated failures at the same site at the same spot is a radius problem or a signal problem, not a discipline problem.
- Re-check after construction, scaffolding or a new adjacent building. Urban GPS accuracy changes when the skyline does.
The operational rule that saves the most grief: treat an out-of-fence check-in as a flag, not a block, at least during rollout. Let the guard mark attendance, record that it was outside the boundary, and put it in front of a supervisor. Blocking creates a guard standing at a gate unable to start their shift, which is a worse operational problem than the one being solved.
What Happens When There Is No Signal
This is the question that decides whether a system works in the field, and it deserves a straight answer.
Two different things can fail independently:
GPS can fail. Underground car parks, inside plant rooms, deep inside steel-framed buildings. The phone cannot establish a position at all, or reports one with an accuracy radius of several hundred metres.
Data connectivity can fail. The phone knows exactly where it is but cannot reach the server.
The second is much easier to handle and most decent systems handle it: the check-in is captured on the device with its coordinates and timestamp, held locally, and synced when connectivity returns. AVES offers limited offline capability of this kind — core field actions can be captured without a connection and synchronise once it is restored, though full functionality does require internet.
The first is harder and there is no clean software fix. If a guard’s post is genuinely inside a signal dead zone, the realistic options are to place the check-in point at the entrance rather than the post, or to accept a flagged low-accuracy check-in and review it. Any vendor claiming their app works perfectly with no GPS is describing something other than geofencing.
Ask this in a demo: what does the app do when GPS accuracy is reported as ±300 metres? A well-built system records the accuracy figure alongside the coordinates so a supervisor can judge the record’s weight later. A poorly built one stores the coordinates as if they were exact.
The Failure Modes Nobody Warns You About
Being honest about limitations is not a weakness in a buying decision — it is the thing that stops a rollout failing in month three.
Device handover defeats it
Covered above, and worth repeating because it is the most common real-world circumvention. Location verification without identity verification is a partial control.
Location spoofing exists
Apps that report false GPS coordinates are freely available. This is a real risk on a fleet of personal Android devices and a much smaller one on managed devices with the relevant settings locked down. Mature systems detect and flag mock-location signals. Ask whether yours does.
Battery and permissions
An app that cannot access location cannot verify anything. Guards turning off location services to save battery, or declining the permission during install, produce silent gaps. Whoever administers the system needs a view of which devices are reporting properly, not just which guards checked in.
Personal devices carry obligations
If guards use their own phones, you are collecting location data about workers on hardware they own. That brings consent, data-retention and privacy duties that vary by jurisdiction and are not optional. At minimum: collect location only at check-in and check-out rather than continuously, tell people plainly what is captured, and set a retention period. Continuous background tracking of staff is a materially different proposition, legally and culturally, from a location stamp at the start of a shift — do not let a vendor blur the two.
It does not fix a supervision problem
If nobody reviews exceptions, geofence attendance produces a very well-organised archive of problems nobody acted on. The technology changes what is knowable. It does not change what gets managed.
Attendance Data as Billing Evidence
The commercial argument for this is usually stronger than the security one, and it is the argument that gets budget approved.
Guarding contracts are priced in deployed hours. The invoice asserts a number of hours; the client either trusts it or does not. When the relationship is good, nobody looks closely. When a contract goes to renewal, or a client’s own auditor gets involved, or there is an incident during a period when the site was supposedly manned, that assertion gets tested.
An attendance record that carries a timestamp, a location and a roster reference survives that test. A muster roll does not.
Three specific places this shows up:
Renewal negotiations. A client asking for a discount is much harder to argue with when neither side can establish what was actually delivered. Verified attendance data changes the conversation from impressions to hours.
Deduction disputes. Clients deduct for unmanned hours. Whether that deduction is fair is answerable in minutes with exception reports, and unanswerable without them.
Tenders. Increasingly, buyers of guarding services ask how attendance is verified. “Paper register” is a weak answer against a competitor who can show a client portal.
There is also an internal case. Payroll built from verified check-ins removes the reconciliation work between what supervisors report, what guards claim and what gets paid — and removes the quiet overpayment that reconciliation usually hides.
Our piece on security guard management software cost covers how to think about the spend side of this.
Rolling It Out Without a Mutiny
Attendance systems fail on adoption far more often than on technology. A rollout that treats guards as the problem produces guards who treat the system as the enemy.
Say what is tracked and what is not. The single most effective thing you can do. Guards assume the worst — that the company watches them all shift. If the system captures location only at check-in and check-out, say so explicitly, in writing, in plain language. If it captures more than that, say that too, because they will find out.
Explain the upside honestly. Verified hours cut both ways. Guards who work a full shift and get shorted on pay have a record now. Guards blamed for a colleague’s absence have a record. Overtime that used to be argued about is now documented. This is a genuine benefit and it is worth leading with.
Start with flag-not-block. As above. The first two weeks should generate data about your radius settings, not disciplinary cases.
Fix the tech complaints fast. The first guard who cannot check in and gets told “it works fine for everyone else” will tell everyone else. Treat early failures as configuration bugs, because they usually are.
Do not bolt discipline onto week one. Introduce the system, let the data stabilise, correct the settings, and only then start managing against it. Operations that lead with enforcement get a fortnight of compliance followed by inventive workarounds.
Brief supervisors properly. They will be asked every question guards have. A supervisor who does not understand the radius or the offline behaviour will improvise an answer, and it will be wrong.
A Worked Example: What a Week of Exceptions Looks Like
Abstract benefits are unconvincing. Here is the concrete shape of what changes, using an illustrative week at a mid-sized contract — the specifics are made up to show the mechanism, not drawn from a real client.
A supervisor opens the exception report on Monday morning. It shows five things from the previous week:
Three late check-ins at the same site, all between 06:35 and 06:50. Not a discipline problem — a bus timetable problem. The 06:00 shift start does not match the only realistic way for guards to reach that site. The fix is a roster change, and it was invisible for as long as the register said 06:00.
One out-of-fence check-in, 400 metres away. The guard checked in from the car park across the road because the gate is inside a covered loading bay with no GPS. This is a configuration finding: move the check-in point or widen the fence for that site.
One missing check-out on a Thursday night. The guard finished and went home without closing the shift. Payroll would have paid the rostered hours; the record now shows the shift was never formally closed. Whether that matters depends on your rules, but you know it happened.
Two check-ins with reported GPS accuracy worse than 200 metres. Both at the same tower block. Not fraud, physics — and a signal that this site’s records carry less evidential weight than others, which is worth knowing before you rely on them in a dispute.
One check-in at 05:12 for an 06:00 shift. Almost certainly a guard arriving very early rather than anything sinister, but it is the kind of pattern that, repeated across one individual, is worth a conversation.
Notice what this list is mostly made of. One possible integrity issue, and four operational or configuration findings that improve the system rather than punish anybody. That ratio is typical and it is the argument to make internally when someone frames this as surveillance. Most of what geofence attendance surfaces is bad configuration and bad rosters, not bad guards.
Notice also that none of it required anybody to watch anything in real time. The value came from a report somebody read on a Monday.
What Attendance Disputes Actually Cost
It is difficult to give a credible figure here and anyone who quotes you an industry-wide percentage is guessing, so instead here is the structure of the cost, which you can populate with your own numbers.
Unworked hours billed. Take your average shift length, your hourly bill rate, and an honest estimate of how often shifts start late or end early without being recorded. Multiply across your guard count and a year. Most operators who do this arithmetic for the first time find the number uncomfortable, because a fifteen-minute average discrepancy is invisible per shift and substantial per annum.
Reconciliation labour. Count the hours per month your admin team spends chasing supervisors for muster rolls, resolving contradictions, and rebuilding timesheets. That is a direct, recurring, fully loaded salary cost.
Deductions and credits. What you gave back last year in client credits for coverage disputes you could not disprove.
Renewal risk. The hardest to quantify and usually the largest. A contract lost partly because you could not evidence delivery costs a multiple of everything above.
Run those four and you have a defensible business case built on your own operation rather than a vendor’s brochure.
Where AVES Fits
AVES records geo-verified attendance as part of the same platform that handles patrol tracking, incident reporting, visitor and gate pass records, shift scheduling and audit-ready compliance reporting — one dashboard across one site or many, rather than an attendance tool that has to be reconciled against everything else at month end.
Practically, that means attendance sits next to the roster it should be measured against, and next to the patrol and incident data from the same shift. When a client asks what happened on a given night, the answer comes from one place.
AVES offers a free trial and a personalised demo. If you are evaluating this, the questions worth bringing are the ones in this article: what the radius defaults are, what happens at ±300m accuracy, whether mock locations are flagged, and what identity check sits alongside the location check.
Explore AVES Plans · Learn More About AVES
For the wider picture, see the AVES Security Guard Management System guide, and security guard scheduling software for the roster side that attendance is measured against.
Frequently Asked Questions
What is geofence attendance?
Geofence attendance is a clock-in method that accepts a guard’s check-in only when their phone’s GPS position falls inside a virtual boundary drawn around the site. The stored record carries a time, a location and a roster reference rather than just a signature.
How accurate is geofence attendance?
It is as accurate as consumer GPS, which typically means a few metres in open conditions and considerably worse among tall buildings, under cover or in bad weather. This is why the fence radius has to be set to suit the site, and why systems should record the reported accuracy alongside the coordinates.
Can guards cheat a geofence attendance system?
Partially. A geofence proves a device was at the site, not which person was holding it, so handing a phone to a colleague defeats it. Mock-location apps can also falsify coordinates on unmanaged devices. Pairing location with a photo capture at check-in and flagging mock-location signals closes most of this gap.
Does geofence attendance work without internet?
Check-ins can generally be captured offline and synced when the connection returns — AVES supports this for core field actions, with full functionality requiring internet. A total GPS blackout, such as an underground car park, is a different problem and needs the check-in point moved or the low-accuracy record flagged for review.
Is geofence attendance legal for tracking employees?
Rules vary by jurisdiction, so this is not legal advice and a local check is worth doing. The general principle is that capturing location at check-in and check-out for a legitimate purpose, with staff informed and a defined retention period, sits on much firmer ground than continuous background tracking of workers.
What is the difference between geofence attendance and patrol tracking?
Attendance establishes that a guard arrived, when, and how long they stayed. Patrol tracking establishes that they covered the site during the shift through checkpoint scans. They answer different questions and a guard can satisfy one while failing the other.
What radius should a geofence be set to?
There is no universal number, because it depends on the site’s size and GPS conditions. Start wider than feels correct, watch the rejection log for a fortnight, and tighten from observed data — a fence that rejects legitimate guards will be worked around rather than complied with.
