Geofencing

How it works

A geofence is a point and a radius, sometimes a polygon, plus an event: entry, exit, or dwell for a set duration. The operating system does the monitoring, not your server. On iOS, CoreLocation region monitoring wakes an app when a boundary is crossed and allows 20 monitored regions per app, which is a real constraint for a company with 60 sites. Android's geofencing API, part of Google Play services, allows up to 100 per app and batches events to save power.

Permissions decide what's actually possible. Both systems require background location, which means the Always Allow authorization on iOS, where the OS periodically reminds the user it's running and the user can revoke it at any time. Android split background location out as a separate permission in Android 10 and pushed it behind a settings screen in Android 11, then added the coarse versus precise choice in Android 12. On a fully managed Android device the management agent can grant that permission silently; in a work profile, the personal side stays invisible either way.

Worth being precise about what an MDM itself can do. The Apple MDM protocol has no geofence payload: geofencing on iOS needs an app holding location permission. The one location command in the protocol is tied to Lost Mode on a supervised device, where an administrator can ask the device for its position after marking it lost. Android offers more through a device policy controller, but the trigger logic still lives in an app.

Accuracy and battery are a trade-off you can't wish away. GPS gives 5 to 20 meters outdoors and effectively nothing inside a steel-framed warehouse, where Wi-Fi and cell positioning fall back to 50 to several hundred meters. A 100 meter fence around a small depot will produce false exits all day. Pad the radius, require a dwell time before acting, and accept latency: the OS APIs save power precisely because they don't poll, so an exit event can arrive minutes late.

Why it matters for a fleet

The convincing use cases are all shared or dedicated hardware. A tablet that only leaves kiosk mode inside the store. A rugged scanner that raises an alert when it crosses the yard gate, which is how most of them get recovered. No personal device, no individual, no history of anybody's movements.

On employee phones, the legal picture in the EU is the deciding factor, not the technology. Location data is personal data: you need a legal basis under Article 6 GDPR, and consent from an employee is rarely valid given the power imbalance. France's CNIL has held for years that continuous geolocation of employees is disproportionate outside narrow situations such as transporting goods or people, lone worker safety, or billing a service, and that it must not run outside working hours. Systematic monitoring calls for a data protection impact assessment, France requires informing and consulting the works council, and in Germany a works agreement is effectively mandatory. Be honest with whoever asks for the feature: tracking staff phones continuously is constrained, and a request framed as security often turns out to be a management request in disguise.

How Appaloosa handles it

Appaloosa doesn't sell employee location tracking, and for most of the needs behind a geofencing request it isn't what you need. Compliance policies are evaluated continuously on device state (OS version, passcode, encryption, required apps, last check-in) rather than on coordinates, kiosk mode locks a shared device to the apps it's meant to run, and remote wipe or selective wipe handles a device that left the site for good. That combination covers the shared-device scenarios without building a movement history of your staff. See the mobile device management scope for what's enforced on enrolled devices.