What is a good sleep score? Why Oura, WHOOP, Garmin, Fitbit and Apple disagree

You wake up feeling great. Your ring says 68 and suggests you take it easy. Your partner slept worse by their own account, tossing at three in the morning, and their watch gave them an 84. By breakfast the two scores have started an argument nobody can win.

Nobody can win it because a sleep score is not a measurement. It is a brand’s opinion of your night, on a scale it chose, against targets it set. Here is what each brand counts as good, and why the same night lands differently on each.

What counts as a good sleep score on each device

Five of the six big platforms score out of 100. That shared scale is the trap, because it makes the numbers look interchangeable when the bands, and the ingredients, are not.

Read the middle column. An 80 is a good night on Oura, Fitbit and Garmin, and only OK on Apple, which saves High for 81 to 95. WHOOP is not on the same scale at all: its percentage is how much of tonight’s sleep need you met, and that need grows after a hard day.

The averages are lower than the bands imply, too. Google says most Fitbit users average between 72 and 83, and Garmin found its users averaged 72 across a full year, with only 7 percent averaging excellent. If your scores live in the 70s, you are sleeping like most people with a tracker, not failing at it.

Why the same night scores differently

Three things pull the scores apart, and none of them is your sleep.

Different ingredients. Apple’s score is mostly about time: duration is half the points, and bedtime consistency over the last 13 nights is almost a third. Oura and Garmin lean much harder on sleep stages, restlessness and, for Garmin, your stress level during the night. Feed the same night to both recipes and you get two different dishes.

Different targets. Google compares your night with targets tailored to your age, sex and time asleep. WHOOP’s target moves nightly: more strain means more sleep needed, sleep debt raises it and a nap lowers it. So a seven-hour night can be plenty on Tuesday and short on Saturday after a long ride.

Stages are the weakest link. Every tracker guesses sleep stages from motion and heart rate rather than brain waves. A 2025 study of six wearables against lab polysomnography found all of them caught more than 90 percent of sleep, but picked up wake far less reliably, and agreed with the lab on stages only fairly to moderately. A score that leans on deep and REM sleep inherits that uncertainty.

• Duration-led score
• Stage-led score
• Percent of sleep need

Illustrative shapes, not measured data. Seven nights.

The Quantified Scientist runs some of the most systematic public tests of how well these devices track sleep, and his rankings are a useful reality check before you trust any single score:

What to track instead

The score is a summary. The things underneath it are where the useful information lives, and two of them are real units.

• Hours asleep. The American Academy of Sleep Medicine and the Sleep Research Society recommend seven or more for adults. Time asleep still differs a little between trackers, but it means the same thing on all of them.
• Bedtime consistency. Apple gives it nearly a third of its score for a reason. Going to bed at roughly the same time is one of the few sleep levers fully in your control.
• Your own average. Whatever device you wear, the question is whether this week is above or below your normal, not whether the number clears a band someone else drew.

That is the line Fit, the cross-brand fitness competition app, takes with sleep, and it is our app, so read this with that in mind. Fit reads sleep from whichever tracker you wear and shows last night against your own normal. On a crew’s board, sleep appears as health context in hours asleep, never as anyone’s score, with a note that trackers stage sleep with their own models and close gaps are a tie. And if your crew wants a sleep challenge anyway, a Fit sleep contest runs on average hours asleep against a goal you set together, not on any brand’s grade. Sleep sits alongside HRV and resting heart rate in the same family of numbers: read against yourself, never against a friend.

What to do with this

1. Learn your own device’s bands. An 80 is good on most scales and only OK on Apple. Know which scale you are reading before you worry about it.
2. Compare with your own two-week average. A score below your normal is information. A score below someone else’s is not.
3. Work the levers, not the grade. Hours asleep and a steady bedtime move every brand’s score, so they are the two things worth chasing.
4. Never compare sleep scores between people, or between devices. The recipes, targets and stage models all differ. The same is true of every recovery and readiness score built on top of them.