Every tracker on the market promises to improve something. Garmin is built around sport, Oura around sleep, Welltory around recovery. Between them they will count your cadence, your REM minutes and your resting heart rate to one decimal place.
Not one of them will tell you what to do about any of it tomorrow morning.
The reason is not technical
Open the Garmin app and you can put your running cadence and your elevation gain on the same graph. Fine. Now what — which of the two do I change on Saturday, and by how much?
The data is there and the answer is not, and sensors have nothing to do with it. A wristband is not a certified medical device. The moment it says “that chest tightness is yesterday’s session, your numbers look normal, go back to sleep”, it is giving medical advice — and one morning it will be wrong, and the lawyers will arrive. So every one of these companies sits on the same contradiction: they hold the data that would support a real recommendation, and they cannot say it out loud.
What they do instead is work around it. I count three ways.
1. Show you the average and let you do the maths
Garmin tells me I got 3 hours 8 minutes of REM last night, and it tells me what the normal range is. Whether my 3 hours 8 minutes is a problem is left as an exercise for the reader.
This is the cheapest workaround, and the most honest one. Nobody promised you a recommendation. You were handed a ruler and left to measure yourself.
2. Turn the advice into an article
The second way is to publish the recommendation as content, so that it counts as education rather than instruction. Write one piece on what short sleep does to the body and one on what long sleep does, then show the reader whichever one matches last night’s numbers.
Welltory does this well — the feed rearranges itself around your metrics, so it reads as though it were written for you this morning.
3. Invent a metric only the company can read
The third way is the most elegant of the three. Garmin’s Body Battery is a single number out of 100 that rises while you rest and falls while you train, while you are stressed, and after a bad night. Alcohol does not feed into it directly — it wrecks your heart rate variability overnight, and heart rate variability is what the number reads. Garmin’s own guidance is simple enough: a high number means you have the energy to train, a low one means it may be a good idea to rest.
The formula behind it is not public, and that is not an oversight — it is the whole mechanism. A hidden number lets Garmin push you towards a better score without ever naming the thing you are supposed to fix. You get a direction and no instruction, which means there is nothing to hold anyone to afterwards.
Garmin does name the ingredients — heart rate variability, stress, sleep, activity. What it does not publish is the recipe: how much each one is worth. That gap is where the advice lives. The number tells you to go up. It never tells you which lever raises it, or by how much. You are left with a direction and no instruction, and a direction is not something anyone can be held to afterwards.
What it adds up to
None of the three is a personal recommendation, whatever the marketing calls it. The first hands you a ruler, the second hands you an article, the third hands you a score.
In the long run the winner will be whoever finds a way to give a real recommendation and survive giving it. I have a rough guess where the solution may lie. I will later share my thoughts on this theme.
What is your experience with smart trackers? Do they have any effect on your life?