GPS tracker for an outdoor cat: does it actually tell you where she is, or just where she was an hour ago?

The reviews contradict each other because two completely different products are sold under the same words, and the listing rarely says which one you are buying.

Radio direction finders. A small tag and a handheld receiver you carry. No network, no subscription. The handset tells you a direction and a rough strength, and you walk until it gets stronger. Tiny, very light, battery measured in months.

Cellular trackers. A device with its own mobile connection that reports a position to an app. This is the one that shows a dot on a map, and it is the one that needs a subscription, because you are paying for a SIM. Heavier, and battery measured in days.

Now your first question answers itself, and the answer is the opposite of what the marketing implies. For actually finding a cat that has not come home, the radio one is usually better. A map dot puts you within a rough radius of where the device was when it last had signal - and a stuck or hiding cat is typically under a shed, in a garage, or somewhere that radius is fifty metres of dense back gardens you cannot walk into. You end up standing in a lane looking at a dot. The direction finder keeps working at exactly that point, gets stronger as you close in, and does not care about signal.

Where the cellular one wins is the other question: where does she go, over weeks. History, ranges, which gardens, whether she is being fed three streets away. That is genuinely interesting and it is a different purchase.

I ended up with one of each and the honest split is: the radio tag lives on the collar permanently, the cellular one comes out when I want the data.

On weight, since this is where cats and dogs genuinely differ and the listings blur it.

A tracker sized for a dog is a meaningful load on a cat, and the problem is not only the mass. A heavier unit swings, catches on things, and sits badly on a lightweight collar, which is the thing that makes a cat work at the collar until it comes off. The tracker you lose is not the one that failed, it is the one she removed in a hedge on day nine.

What matters in practice:

• Weigh the whole assembly, tracker plus collar plus any attachment, not the figure on the box. The mounting hardware is often a surprising share of it.
• The collar has to be a proper safety-release type for a cat that goes outdoors and climbs. That is not negotiable, and it means anything on it will occasionally be lost with the collar: so buy on the assumption that you will replace it at least once, and check whether spare collars are sold separately and at what price.
• Check the profile, not just the grams. A flat unit that sits against the neck behaves completely differently from a chunky one that hangs and swings, at the same weight.

A practical trick before you spend anything: tape a coin or two of roughly the right mass to her existing collar for a week. If she is at it constantly or the collar keeps coming off, you have learned that for nothing.

The running cost is where people get caught, and it is worth doing the arithmetic before rather than after.

Add up: the device, the subscription for twelve months, at least one replacement collar, and - if it is cellular - the fact that a subscription lapsing usually means the device stops reporting entirely rather than degrading to something lesser. Check that last one specifically before buying, because a tracker that becomes a paperweight the month you forget to renew is a different proposition from one that keeps working locally.

The thing that actually decides whether you still use it in month three is charging. A cellular tracker with live updates wants charging every few days. That means catching the cat, removing the collar, remembering to put it back on. Miss it twice and the device is flat in a drawer on the one night she does not come home, which is precisely when you needed it.

So be honest about the update interval you will really live with. Most of these let you slow the reporting rate down, and the difference between updating constantly and updating every few minutes is enormous for battery and almost irrelevant for finding a cat who has been missing since yesterday. Turn it down on day one. The default is set to demo well, not to run for a year.

The device that is charged is better than the accurate one that is flat.

One thing worth doing whichever you buy, and it costs nothing.

Test it before you need it. Give the tag to somebody, have them hide it in a neighbour's garden or a shed with the door shut, and go and find it. You will learn more in twenty minutes than from any review: how the app behaves when the signal is poor, how close you have to get, whether the interface is usable one-handed in the dark, and whether the thing you assumed would work actually does through a brick wall.

The two failure modes people discover at the worst possible moment are that the app needs a login they have forgotten, and that the last known position is a place the cat obviously is not any more. Both are much less alarming when you have already seen them on a Sunday afternoon.

And the boring thing that still finds more cats than any of this: a microchip that is registered to your current phone number. It costs nothing to check and it is out of date far more often than people think.