What is Dwell time?

Dwell time carries two related meanings in geofencing, and they get conflated often enough to be worth separating up front.

As a measurement: how long a device stayed inside a zone. A vehicle dwelt 42 minutes at a customer site. A phone dwelt 7 hours at school.

As a filter: the minimum time a device must remain on the new side of a boundary before the crossing counts. Cross the line for 20 seconds and come back, and with a 2-minute dwell threshold nothing is reported.

The first is analytics. The second is noise control.

Dwell as a filter

The problem it solves is the pass-through. A geofence around a depot will fire for every vehicle that drives past on the adjacent road, clipping the corner of the zone for a few seconds. Those crossings are geometrically real and practically meaningless.

Requiring the device to stay on the new side for some minimum duration filters them out. Set it to two minutes and driving past stops registering, while actually stopping still does.

It also helps with GPS drift, though less than people expect. Drift produces brief excursions, so a dwell requirement suppresses many of them. But drift is not always brief — a device can report a position 40 metres off for several minutes when satellite geometry is poor. A dwell filter long enough to catch that would delay genuine alerts unacceptably.

The cost: latency

Every second of dwell requirement is a second of delay on genuine events. This is the whole trade, and it is not avoidable.

Dwell threshold	Filters out	Costs you
0 (none)	Nothing	Nothing — but every pass-through alerts
30 s	Fast pass-throughs	Half a minute on real events
2 min	Most incidental traffic, brief drift	Two minutes — often too slow for safety use
10 min	Almost all noise	Useless for anything urgent

For safety applications the calculation usually favours a short threshold or none. If a child leaves school grounds, a two-minute delay is two minutes of not knowing, and they can cover a lot of ground. For asset monitoring — a trailer in a yard — a longer threshold is fine, because nothing is urgent and quiet notifications matter more.

What Geoblip does

Geoblip does not use a dwell-time filter for alerting. It attacks the same problem from a different angle, deliberately.

Rather than delaying an alert to see whether a crossing persists, Geoblip alerts on the first genuine exit and then suppresses everything until the device truly returns. Each device-zone pairing holds an alertStatus of ok or alerting; the exit notification fires on the ok → alerting transition, and re-arming requires a complete alerting → ok → alerting cycle. This is described under alert debouncing.

The reasoning: a dwell filter delays every alert to suppress some noise. State-based suppression delays nothing and suppresses repeats instead. For the use cases Geoblip is built around — children, elderly relatives, equipment — the first alert being immediate matters more than the pass-through case being perfectly silent.

The pass-through problem is better solved by drawing zones that do not touch through-roads in the first place. The geofence builder makes that easy to check.

Dwell as a measurement

The analytics sense is useful in its own right. Time spent inside a zone answers questions raw position data does not:

• How long did the technician spend on site?
• Is the vehicle idling at the depot longer than scheduled?
• Did they arrive at school and stay, or arrive and leave again?

This needs position history rather than just current state — Geoblip’s Pro plan retains 24 hours and Business 30 days, which is what makes dwell computable after the fact. See pricing for the tiers, or exit alert for what happens when the dwelling stops.

Last reviewed 7 August 2026.