Know the moment your dog gets out of the garden, rather than finding out when you next look — with a boundary that matches your actual property.
Most pet trackers are built around live tracking: open the app, see a map, watch a dot. That is useful once you already know the animal is missing. It does nothing about the gap between the escape and the moment you notice — which is usually hours, and which is the part that determines how far they get.
A geofence closes that gap. The alert arrives while the dog is still on your street.
Gardens are the strongest argument for polygons over circles anywhere in geofencing.
A circle around your house either leaves parts of the garden outside the zone — so the dog “escapes” by walking to the end of it — or covers your neighbours’ gardens too, so a genuine escape does not register until they are two doors down. Neither is acceptable when the whole value is early warning.
Trace the actual boundary instead. A dozen points following the fence line takes two minutes in the geofence builder and gives you a zone that matches the property. See GeoJSON polygon for how the shape is stored.
There is a genuine tension here. Every other use case says draw the zone generously; for pets, an over-large zone delays the alert you actually want.
The compromise that works:
Small urban gardens are the hard case. A garden 15 metres across is smaller than the accuracy of many trackers, and no configuration fixes that — see GPS accuracy. In that situation the zone is better drawn around the block than the garden: you lose early warning but gain alerts you can trust.
Weight is the first filter. A tracker comfortable on a labrador is unusable on a cat or a small terrier. Check the grams before anything else — an uncomfortable tracker comes off, and a tracker on the kitchen table protects nothing.
Waterproofing is not optional. Dogs swim, roll in wet grass, and stand in rain. IP67 or better.
Battery versus reporting interval. These trade directly against each other. A tracker reporting every 10 seconds gives fast alerts and needs charging every day or two; one reporting every 5 minutes lasts weeks but may let a dog cover several hundred metres before you hear. For a garden boundary, somewhere around every 60 seconds is a reasonable middle.
Attachment. A tracker on the collar goes wherever the collar goes — including off, if the dog slips it. Nothing to be done about that, but worth knowing it is the failure mode.
Worth being honest rather than selling you something that will disappoint.
Cats roam by design, so a geofence around the garden alerts constantly and tells you nothing. Their small size limits you to very light trackers, which means small batteries and less capable antennas. And they spend a lot of time under cars, in sheds, and indoors, where satellite positions are poor or unavailable.
Geofencing works well for a cat kept to a specific area, or a wide zone that flags when they have gone unusually far. For a free-roaming cat in a normal neighbourhood, live tracking is the more honest tool.
You get an exit alert within a reporting interval or two. When the dog comes back through the gate, you get a recovery alert, so if a neighbour returned them while you were driving around looking, you find out without having to check.
Repeat alerts are suppressed — a dog sleeping against the fence will not generate a stream of notifications. See alert debouncing.
Know if your child leaves school grounds during lesson time — without watching a map all day or tracking their every movement.
Get told immediately if someone living with dementia leaves a safe area — including overnight, when wandering is most likely and least likely to be noticed.
Know within minutes if a trailer, generator, or machine leaves the yard — the window that decides whether it is recovered or gone.
Draw the zone, add the device, choose the hours. Seven-day free trial, cancel any time.