GPS Accuracy and Battery Life: Making Geofenced Clock-Ins Work on Real Phones
GPS drift and battery drain worry managers switching to geofenced clock-ins. Here's how to set a fence that holds up on real phones.
On this page
A punch from inside the fence is accepted. A punch from across the road is denied and logged. That is the whole promise of geofenced clock-ins: the record settles the question, instead of a Friday argument about who was where.
But that promise only holds if the fence is built for real phones, on real sites. GPS wanders. Batteries run down. A fence drawn too tight turns a good system into a daily argument, with staff locked out of punches they should have made. The fix is not a stricter fence. It is a fence sized for how GPS actually behaves, and phone settings that let the punch go through.
Why GPS wanders
GPS is an estimate, not a pin dropped on an exact spot. On an open lot, a phone’s reported position is usually close to where the person stands. Inside a warehouse, a basement, or between tall buildings, the signal bounces and the estimate drifts. The phone may report a worker a few dozen metres from where they actually are, even though they never moved.
That drift is normal, and a geofence has to expect it. If the boundary hugs the building walls exactly, drift alone will push legitimate punches outside it. Managers then spend the week fixing exceptions that were never really exceptions, and staff stop trusting the system.
The practical answer is tolerance. A geofence is a boundary plus a margin for error, and the margin should match the site, not the map. As a rule of thumb used across geofencing tools, a radius of roughly 50 to 200 metres covers most sites, with the wider end for large or obstructed ones.
Size the fence for the site
Walk the site before you draw the fence, or at least picture how staff actually arrive. Where do they park? Which door do they use? Where does the signal get weak? Then size accordingly:
| Site type | What to allow for |
|---|---|
| Small shop or office | A tight radius, but still room for drift at the door and the car park |
| Warehouse or yard | A wider radius that covers the yard, loading bays, and the staff entrance |
| Construction or rural site | The widest radius, since boundaries are vague and signal can be patchy |
| Multi-building campus | One fence per building where punches must be tied to a specific building, or one site fence where they do not |
Two mistakes cover most setup trouble. The first is drawing the fence around the building footprint only, so the car park and entrance fall outside it. The second is copying the same radius to every site, so a downtown shop and a sprawling yard get the same margin. Test the fence with a real phone at the real entrance before you roll it out to the whole team. One test punch at each entry point will tell you more than the map will.
Battery: check at the punch, not all day
The other worry managers hear first is battery life, and it is a fair one. Nobody wants a time clock app draining a phone before the shift ends. The reason geofenced clock-ins do not have to do that is simple: location only needs checking at the moment of the punch.
A geofenced punch asks one question at clock-in and again at clock-out: is this phone inside the fence right now? It does not need to follow the phone around for the rest of the shift. That distinction matters twice over. It keeps battery use to a short check instead of continuous tracking, and it answers the surveillance objection honestly. Staff are not being watched all day; a single location point is recorded with each punch, and that record is what settles disputes later.
If staff are worried about battery, that is the explanation to give them first, before any settings talk. Then sort the settings, because a phone trying to save power can also be the thing that breaks a punch.
Phone settings that make or break a punch
On both Android and iOS, two settings cause most failed punches, and both are easy to check during rollout week:
- Location permission: the app needs location access when it is in use, so it can check the fence at clock-in and clock-out. If permission is set to never allow, or to approximate location only, the punch cannot be verified.
- Battery optimisation: aggressive battery savers can delay or block an app running in the background. If punches fail on one person’s phone but work on everyone else’s, battery optimisation on that phone is the first place to look.
- Signal, not just GPS: a phone indoors may lean on Wi-Fi and cell towers to place itself when satellites are weak. A punch taken just inside a thick-walled building can drift further than the same punch taken at the door.
Give staff a one-line version of this on rollout day: keep location on for the app, do not let the phone put it to sleep, and step to the entrance if a punch is refused indoors. Most “the app doesn’t work” reports turn out to be one of those three.
When a punch still falls outside
Even a well-sized fence will refuse a punch now and then. GPS has bad minutes, and a worker will occasionally try to punch from the wrong side of a large site. What matters is what happens next.
Treat the first outside punch as information, not a violation. Check the record: where was the phone, and how far outside the fence? A punch a few metres out at the usual entrance is drift, and the fence or the habit needs adjusting. A punch from a different neighbourhood is a different conversation. Reviewing exceptions before payroll, rather than letting them pile up into a Friday reconstruction, keeps the record honest and keeps staff on side.
A simple weekly habit covers it: review refused and corrected punches once a week, widen a fence that keeps catching legitimate arrivals, and talk to the person before assuming the worst. The fence is there to produce a record everyone trusts, and trust runs both ways.
A rollout checklist you can copy
- Draw the fence around entrances and parking, not just the building footprint.
- Size the radius to the site: tighter for a small shop, wider for a yard or rural site.
- Test one punch at each entrance on a real Android phone and a real iPhone.
- Tell staff plainly that location is checked at the punch, not tracked through the shift.
- Check location permission and battery settings during rollout week.
- Review exceptions weekly, and fix the fence before blaming the punch.
Done that way, geofencing stops being a gadget and becomes what it was meant to be: proof, not trust. The punch carries its own evidence, the battery survives the shift, and Monday’s payroll starts from a record instead of a story. For more on running time across many sites, see the rest of the ClockLap blog.