Geofence Time Clock for Construction Sites: Setting the Right Radius

I have set up hundreds of geofences for construction crews in the last few years, and the question I get every single time is the same: “How big should we make it?”

Too small and your guys can’t clock in from the gate. Too big and the geofence does nothing — somebody can clock in from the diner across the street and the app says, sure, that’s the job site.

This is the practical, field-tested guide to setting a geofence radius that works on the first day, holds up in a rainstorm, and does not turn the foreman into a help-desk for the app.

What a geofence actually does

A geofence is a circle drawn on a map around a job site. When a crew member clocks in, the time tracking app checks the phone’s GPS coordinates against that circle. Inside the circle, the clock-in is allowed. Outside, the app flags it — either blocking the punch outright or recording it as an out-of-fence event that the supervisor reviews later.

Klees combines the geofence check with a PinShot selfie verification, so the record on every clock-in is: who, where, and when, with an anti-spoof-scored photo attached. The geofence is the “where” half of that record. Get it wrong and the whole verification chain wobbles.

Why the wrong radius costs you both ways

A geofence that is too tight is the single fastest way to lose foreman trust in a new app. The crew shows up, lines up at the trailer, opens the app, taps clock-in, and the app says “outside fence.” Now the foreman is on the phone to the office. The office adjusts the fence. The crew waits. Productivity is gone for that morning.

A geofence that is too loose is the opposite problem and the one that gets the owner in trouble. If the fence covers the entire block, then a worker can clock in from a parking lot two streets over and the system records it as on-site. The audit trail is intact but it is recording the wrong thing.

The right radius is the smallest one that still accepts every honest clock-in. That is the rule.

Default radii by site type

After running this exercise across our customer base for construction operators on commercial buildings, single-family residential, road work, and post-construction cleanup, the defaults I recommend are:

Site type	Recommended radius	Notes
Single-family residential	50 m	Tight, because the lot is small. Confirm GPS isn’t bouncing.
Single-building commercial	75 m	The standard default. Covers the gate, the trailer, and the lay-down area.
Multi-building campus	120 m	Use one geofence per building if scopes differ.
Industrial lay-down yard	150 m	Bigger because crews stage equipment outside the structure.
Road and highway work	300 m+ as a moving fence	Linear sites need a different model — see below.
Post-construction cleanup	75 m	Same as commercial. Cleanup crews work the same footprint.

These are starting points. The right number for your site is whatever radius accepts the foreman’s clock-in from where they actually arrive on the job — usually the parking area, sometimes the gate, sometimes the trailer. Stand there with the phone. Look at the GPS. Set the radius.

Where GPS gets weird

GPS accuracy is not a single number. It depends on the phone, the sky view, the buildings around the site, and the weather. In open conditions a modern phone is accurate to 3–5 m. Under a high-rise downtown, you can see accuracy drift to 30 m. Inside a steel-frame building, GPS does not work at all and the phone falls back to WiFi triangulation, which can be off by 100 m.

For construction sites this means:

• Downtown urban sites — bump the radius by 25 m to account for canyon-effect drift.
• Sites surrounded by metal storage — same.
• Sites where crews clock in from inside a partially built structure — same.
• Open sites — the standard radius is enough.

The U.S. government’s GPS performance standard documents the underlying accuracy ranges if you want the long version.

The moving-fence problem on highway and pipeline work

Linear construction — highway resurfacing, pipeline, transmission line — does not have a single point to draw a circle around. The crew might be a mile of road today, two miles tomorrow. A fixed-radius geofence does not fit.

For these sites, the right model is one of two things. Either you redraw the geofence daily based on where the crew is working — which requires somebody at the office to update it — or you use a moving fence that follows the foreman’s truck. Klees supports both. In practice, for linear work, I recommend a daily redraw because it leaves the audit trail cleaner: the geofence is fixed for that shift, the time records reference a known perimeter, and disputes resolve faster.

How to set the radius on day one

The setup procedure that has not failed me in three years of rollouts:

1. Walk the site with the foreman. Phone in hand, app open.
2. Identify the four extreme points crews actually reach during a normal shift — parking, gate, trailer, lay-down area.
3. Drop GPS pins at each one.
4. Set the geofence center at the centroid of those pins.
5. Set the radius to the largest distance from the centroid to any pin, plus 15 m of buffer.
6. Test clock-in from each point. All four must accept.
7. Test clock-in from outside the fence — across the street, at the gas station next door. The app must reject.

If steps 6 and 7 both pass, the geofence is right. Save it. Move on.

What to do when a geofence rejects an honest worker

It will happen. Not often, but it will. GPS drifts. A phone is acting up. A worker is standing exactly on the edge of the fence.

The pattern that works:

• The app permits an override — the worker submits an out-of-fence clock-in with a reason. Klees attaches the GPS coordinate and the PinShot selfie to the record so the supervisor can review.
• The foreman gets a notification. The override is visible immediately in the Live Map view.
• The supervisor reviews at end of shift. If the GPS pin is clearly on-site (the worker was 8 m past the fence because GPS bounced), the supervisor approves. If the pin is across town, the supervisor does not.

The override is the pressure valve. Without it, foremen lose patience with the app fast. With it, the rare false-reject does not derail a shift.

How to audit the geofence monthly

Once a month, walk this list:

1. Look at the count of out-of-fence clock-ins per site.
2. Sites with more than 5% out-of-fence rate need the radius bumped 25 m or the perimeter redrawn.
3. Sites with zero out-of-fence rate over 60 days may be too loose — narrow the radius by 20 m.
4. Review every approved override. If a site has a recurring pattern of overrides at the same coordinate, your fence is in the wrong place.

This 15-minute monthly review is what keeps the system honest. Without it, geofences drift toward the loose end as foremen approve overrides, and within a year the fence is doing nothing.

How geofences fit with PinShot and Live Map

A geofence by itself answers the “where” question on a clock-in. Combined with PinShot (the “who” question, with anti-spoof selfie verification) and the time stamp (the “when” question), you have a complete record. The Live Map view shows every active crew with their on-fence status in real time, so the supervisor sees the whole operation at a glance.

For multi-site operators this matters. A cleaning or construction crew running 20 active sites in a day has an unmanageable problem on paper. On Live Map with geofences set correctly, the supervisor sees one screen.

FAQ

What is the best default geofence radius for a construction site?

75 m is the strongest default for a single-building commercial site. Adjust up for urban canyons (add 25 m), industrial lay-down (use 150 m), or linear work (use a moving fence or daily redraw).

Does the geofence work without cell signal?

The GPS chip works without cell signal — it talks directly to satellites. What requires connectivity is uploading the clock-in to the server. Klees holds the clock-in locally and syncs when signal returns, preserving the original timestamp and GPS pin. See Offline Time Tracking for the full mechanics.

Can workers spoof their GPS location?

In theory, yes, with a rooted device and a mock-location app. Klees flags devices with developer mode enabled and devices reporting impossible velocities (you cannot be at site A and site B 50 miles apart 30 seconds later). When the selfie verification is combined with GPS, a successful spoof requires both a faked location and a real-time photo of the right person — a combination that has not appeared in our customer base.

What happens to off-site work like supply runs?

You either create a small geofence around the supply yard or you allow a “travel” cost code that does not require an on-site clock-in. Most operators choose the second approach.

Does Klees support multiple geofences per job?

Yes. A campus with three buildings can have three geofences. Each can be tied to a different cost code, so when a worker enters Building B, the system can prompt for the right scope.

---

Need help dialing in geofences for a multi-site operation? Book a setup walkthrough with Klees field-ops — we have done this for everything from a 3-house remodel shop to a 400-site cleaning operator.