Introducing It

Geofenced clock-in lands better than most monitoring, because its narrowness is genuinely reassuring — if the narrowness is real and is explained.

Decide these before announcing

Does the app collect location between punches? Decide no, and verify it.

What is the radius, per site, and how was it chosen?

What happens when a punch is refused? This is the question people care about most.

How long are coordinates kept?

Is there a fallback for someone without a working phone?

Will flagged punches be used in disciplinary matters? Decide, and write it down.

Measure before you deploy

Walk each site and record position and accuracy, as described in the accuracy note.

Set radii from those measurements, not from a default.

A deployment that begins with a week of refusals never recovers the goodwill, and the refusals are entirely predictable from the measurements.

What to say

What is checked: your position, once, when you tap to clock in.

What is not: your location at any other time.

How to verify that yourself, through the phone's permission screen.

What happens if it fails: the punch goes through flagged, you start work, it is sorted afterwards, and you are paid.

That last sentence does more for acceptance than everything else combined.

Piloting

One site, willing, for a full month including a busy period.

Measure the refusal rate, which is the number that matters.

Ask what went wrong and believe the answers: the entrance that never works, the basement, the phone model that behaves differently.

Fix the radii before extending.

A pilot that produces no changes was not a pilot.

The first month

Publish the refusal rate, which signals that failures are the system's problem rather than the worker's.

Fix the worst site visibly.

Report what the data showed about the work — a shift start that does not match the rota, a site where everyone arrives early because of a bus timetable.

What undermines it

A refusal treated as lateness.

Corrections that take days.

A radius nobody will explain.

Discovering the app collects more than was said, which is checkable on any phone and will be checked.

Any of these in the first month, and the system is understood as surveillance regardless of what it actually does.

Measure before you announce

The sequence that prevents the worst first week.

Walk each site and record position and accuracy.

Set radii from those measurements.

Then announce.

A deployment that opens with a week of refusals never recovers the goodwill, and the refusals were entirely predictable from measurements nobody took.

Check the difficult case

Use the scheduling example to frame one representative test. The useful evidence is the record created when an employee corrects an entry and an administrator exports it.