Geofence Attendance: How to Choose the Right Radius

A geofence for attendance is a virtual circle around a workplace: a clock-in only counts when the phone is inside that circle. Make it too small and people standing at the gate get rejected; make it too large and someone in a café 500 m away can clock in. This guide gives starting radiuses by site type, explains why GPS accuracy matters more than the radius number itself, and shows how to handle the awkward cases at the edge.

How a geofence decides “inside”

Each site has a center (latitude and longitude) and a radius in metres. At clock-in, the system measures the distance from the phone’s position to the center. If that distance is smaller than the radius, the person is inside.

The catch is that a phone never reports an exact point. It reports a point plus an accuracy estimate, often shown as “±20 m”. Under open sky the error can be a few metres. Inside a concrete building, in a basement or between towers it can reach hundreds of metres. Our article on phone GPS accuracy explains why.

A sensible radius therefore accounts for:

The real footprint of the working area.
The typical GPS error at that location.
Where people actually stand when they clock in: the gate, the car park, the locker room.

Starting radiuses by site type

Treat these as starting points, not standards. Measure during your pilot week and adjust.

Site type	Starting radius	Why
Office in a high-rise (District 1, Cau Giay)	150–250 m	Indoor and urban-canyon GPS drifts a lot; staff may be on high floors or in basement parking
Shophouse office or retail store	100–150 m	Small space, but you still need room for error and for people at the entrance
Construction site	Half the plot’s diagonal + 50–100 m	Large plots with site huts and material yards at the edge
Warehouse or factory in an industrial park	200–400 m	Big footprint; metal roofs and steel frames degrade GPS
Retail outlet or supermarket (field sales)	100–200 m	Multi-storey malls produce large indoor errors
Residential compound or resort (security, cleaning)	Several smaller sites, or 300–600 m	One huge circle would swallow surrounding neighbourhoods

One simple rule: avoid going below 100 m unless you have confirmed accuracy at that spot is consistently good. Most “I was standing at the gate and it said outside” complaints come from radiuses that are too tight.

Multi-site and multi-branch teams

If you operate several branches, create one site per location rather than one giant circle. You gain:

Timesheets that show where someone clocked in, not just whether they did.
Different radiuses for a small office and a large warehouse.
Clean closure: archive a site when a project ends.

When two sites sit close together, such as two towers in the same compound, their circles may overlap. That is fine, but check which site the record is assigned to. A common approach is to pick the site the person is most deeply inside.

Edge cases you will meet

Indoor drift

Someone is at their desk but the phone places them 300 m away. Allow some margin for GPS error, encourage staff to keep Wi-Fi on (it helps positioning), step near a window, and have a process for admins to approve exceptions.

Standing right on the boundary

The position wobbles by ±15 m and the status flips between inside and outside. A good system adds a small delay before changing status so it does not flicker. On your side, the easy fix is to widen the radius by a few dozen metres.

Very weak GPS

When the accuracy estimate is very poor (over 100 m), you cannot honestly say whether someone is inside. You can block the clock-in (strict but frustrating) or allow it with a flag for later review. For most SMEs, the second option works better.

Work at a client’s location

Staff sent to a client site that is not in your list need either a temporary site or permission to clock in outside the area with a flag for that period.

Common configuration mistakes

Using the map address as the center instead of the real gate or work area. New construction plots are often mislocated on maps.
One radius for every site.
No field test: send someone to walk the boundary and take test photos at the gate, car park and basement.
Forgetting that a geofence does not prove identity: it tells you where the phone is, not who is holding it. See how to stop attendance fraud.

Geofences in SnapID Mark

In SnapID Mark, admins set up sites (center and radius) in the web dashboard. In the camera, staff immediately see a status chip, “Inside · site name” or “Outside · X from site”, along with the current GPS accuracy.

The company chooses what happens outside a site: block the photo, or allow it and flag it as outside the geofence for review.
When GPS is weak, the app shows a “Weak GPS” warning and waits briefly for a better fix; if it does not improve, the photo is taken with a flag instead of leaving the person stuck.
The app allows for GPS error and avoids flickering at the boundary.
The server recalculates the geofence for every upload; the phone’s result is only an initial signal.
The app does not track location in the background: location is used only while the camera is open.

See the security and business sections for more.

FAQ

What is the smallest radius I should use?

For most sites, 100 m is a safe minimum. Smaller circles tend to cause false rejections from GPS error, especially indoors.

Can I use a polygon instead of a circle?

Some systems support polygons. SnapID Mark uses circles (center plus radius); for a long, narrow area, place several circular sites side by side.

Will someone on the 30th floor show as outside?

Possibly, because GPS inside towers is weak. Widen the radius for that site, encourage Wi-Fi, and let people clock in at the lobby if needed.

Does a geofence mean staff are tracked all shift?

It does not have to. In SnapID Mark, location is captured only when a photo is taken, never continuously.

Want to test it on your own sites? Download SnapID Mark, create one site and walk its edge to find the radius that fits.