A GPS dog fence has no buried wire. You draw a boundary in an app, the collar reads its own GPS position, and when the dog crosses the line the collar warns and then corrects. It is a genuinely clever product and it has one hard limitation that the marketing does not lead with.
The canopy problem, stated plainly
The boundary is only as accurate as the collar's GPS fix. Under open sky a consumer GPS fix is good to a few meters. Under heavy tree canopy — mature woodland, dense conifers, a garden with large trees — accuracy degrades to tens of meters.
A boundary that wanders by twenty meters is worse than no boundary at all. It corrects the dog while it is lying in its usual spot, and fails to correct when the dog crosses the real line. The dog learns nothing except that the collar is unpredictable, which is both a training failure and a welfare problem.
So: look up at your own trees before you spend five hundred dollars. On open land these systems do what they claim. Under canopy, be skeptical of every brand equally — this is physics rather than a quality difference.
What you are actually choosing between
A correcting fence. SpotOn and Halo. The collar delivers a warning tone and then a correction. This is the only option that actually contains a dog without a physical barrier, and it needs open sky.
An alerting geofence. Tractive and most cellular trackers. No correction — the collar simply tells your phone the dog has left the area you drew. For a lot of gardens this is what people actually wanted, it is far cheaper, and it has no welfare question attached.
Neither: a findable dog. If containment is unreliable on your land, the honest alternative is to accept that the dog will sometimes leave and make sure you can find it. A radio system works under canopy where GPS containment does not, and has no subscription.
Questions to ask before buying any wireless fence
What does my tree cover look like?
The single most important question and the one the sales process never asks. Open grassland: fine. Scattered mature trees: variable, and test the boundary carefully. Dense woodland: do not buy a GPS fence.
What is behind the boundary?
A GPS fence is not a barrier — it is a deterrent that depends on the dog respecting it. If twenty meters of drift puts the dog on a road, the consequence of a bad fix is not an inconvenience. Wireless containment is for land with forgiving edges.
What happens when the collar battery dies?
The fence stops existing. A physical fence does not have a battery. This is worth building a routine around rather than discovering once.
Which features are behind a subscription?
Halo in particular tiers its features, and the tiers include things people reasonably assume are part of the hardware. Read the tier list before buying, not after — a fence whose boundary-count or alert features lapse with a payment is not the product you thought you bought.
Will this dog respect a boundary at all?
Some dogs will not. A high prey-drive dog mid-chase does not care about a tone, and a dog that breaches once and is not corrected learns that breaching works. If your dog has already defeated a physical fence, a virtual one is unlikely to succeed.
The honest comparison with a buried wire
A traditional buried-wire fence has no GPS accuracy problem at all, works identically under canopy, and costs considerably less than a premium GPS collar. What it costs you is a weekend with a trencher and permanence: the boundary is where you dug it.
The real advantages of GPS are flexibility — you can redraw the boundary, and you can create one away from home — and installation, which is zero. If you will use the flexibility, GPS is worth it. If you want one boundary around one garden forever, a wire is the more reliable and cheaper technology, and we would rather say so.
Training matters more than the hardware
Every manufacturer says this and it is true. These systems work when the dog has been taught the boundary with flags and supervision over a couple of weeks, and they fail when a collar is fitted and the dog is let out.
If you are not going to do the training program, do not buy a correcting fence. Buy an alerting geofence instead — it asks nothing of the dog, it tells you when it has gone, and nobody gets corrected for something they were never taught.
Where we land
On open land, with forgiving edges, and an owner who will do the training: a GPS fence is a good product and SpotOn is the one we would look at first.
Under trees, near a road, or without the training time: do not buy one. Buy a geofencing tracker that alerts you, or a radio system that finds the dog, and spend the difference on a gate latch.
