Consumer sleep trackers are genuinely useful for one thing and unreliable for another, and most people have it backwards. They are good at telling you when you slept and roughly how long. They are poor at telling you which sleep stage you were in, which is precisely the number people fixate on. And in a minority of users, the tracker itself becomes the problem: chasing a better sleep score produces exactly the performance anxiety that keeps people awake. There is a clinical name for that now.
Key takeaways
- Trackers estimate sleep from movement plus heart rate. They do not measure brain activity, which is how sleep stages are actually defined.
- They are reasonably good at total sleep time and timing, and considerably weaker at stage breakdowns.
- They tend to overestimate sleep in people who sleep badly, because lying still and awake looks like sleep to an accelerometer.
- No consumer device diagnoses sleep apnoea. A flag from a watch is a prompt to get tested, not a result.
- Use the trends and ignore the nightly score. If the score is making you anxious, stop looking at it.
How they actually work
A sleep laboratory scores sleep stages from brain activity, eye movement and muscle tone. That is the definition of a sleep stage: it is an electrical pattern.
Your watch or ring has none of those sensors. It has an accelerometer measuring movement, and an optical sensor measuring heart rate and heart rate variability. From those two signals, an algorithm infers whether you are asleep and which stage you are probably in.
That inference is reasonable for the broad question, asleep or awake, because sleep involves much less movement than wakefulness. It is far shakier for the finer question, because REM, light and deep sleep are not reliably distinguishable by movement and heart rate alone.
So the hierarchy of trust runs roughly like this:
| What it reports | How much to trust it |
|---|---|
| When you went to bed and got up | Good |
| Total sleep time | Reasonable |
| Consistency of your schedule across weeks | Good, and the most useful thing it does |
| Number of awakenings | Moderate |
| Time in each sleep stage | Weak |
| The overall "sleep score" | A proprietary composite. Treat as a rough trend |
The overestimation problem
Trackers tend to overestimate sleep in people who sleep badly, and the reason is structural rather than a bug.
If you are lying still in the dark, awake, frustrated, your movement profile looks a great deal like sleep. Your heart rate is low. You are not moving. An algorithm built on those two signals will often score that as sleep.
The irony is sharp. The people most likely to buy a sleep tracker to investigate their poor sleep are the people whose poor sleep it is least able to see. Someone with insomnia may be told they slept seven hours on a night they experienced as three hours of frustrated wakefulness, and both accounts are, in their own terms, correct.
Orthosomnia
In 2017, sleep clinicians described a pattern they were seeing often enough to name: patients whose preoccupation with achieving perfect sleep-tracker data was making their sleep worse. They called it orthosomnia.
The mechanism is straightforward once you see it. Sleep is one of the few things that reliably retreats from effort. Introduce a nightly score, and you introduce a performance target. Now bedtime carries an evaluation, you are monitoring yourself for signs of progress, and you check the number the moment you wake. That is a recipe for exactly the arousal that maintains insomnia, and we describe the same loop in racing thoughts at night.
Some people also start adjusting their behaviour to please the algorithm: going to bed earlier than they are sleepy to bank more hours, which weakens the association between bed and sleep and makes things worse.
This is a named clinical observation from a case series rather than a measured population phenomenon, so we are not going to tell you how common it is. But if you recognise yourself, the intervention is simple and it works: stop looking at the score.
What a tracker is genuinely good for
Consistency. This is the best thing these devices do, and almost nobody uses them for it. Your bedtime and wake time across four weeks, laid out visually, is genuinely informative, and schedule regularity is one of the strongest levers you have over sleep quality. If your weekend wake time is three hours later than your weekday one, your tracker will show you that, and it matters more than any stage breakdown.
Timing patterns. If your data shows you consistently falling asleep at 2am and struggling every weekday morning, that is a real signal, and it points toward a body-clock question rather than an insomnia one. See night owl or delayed sleep phase.
Trends over weeks. Direction of travel over a month is far more meaningful than any single night. Single nights are noisy for everyone, including good sleepers.
Prompting you to act. A device flagging irregular breathing or low overnight oxygen has done something useful if it sends you for proper testing.
What it cannot do
It cannot diagnose sleep apnoea. Some devices now flag possible breathing disturbances, and that is a reasonable prompt. It is not a diagnosis, and a device that says nothing does not exclude the condition. Diagnosis requires measuring airflow and respiratory effort, which is what a sleep study does.
It cannot diagnose insomnia. Insomnia is a clinical diagnosis based on your experience, its duration and its daytime consequences. A device telling you that you slept seven hours does not overrule your experience of the night.
It cannot tell you your sleep stages with real confidence. So the anxiety about not getting enough deep sleep is usually anxiety about a number that is an estimate produced by an algorithm from a wrist-based heart rate.
It cannot tell you why. This is the fundamental limitation. Even where the data is accurate, it describes what happened, not the cause. Fragmented sleep looks similar whether it comes from apnoea, alcohol, pain, a mood disorder or a noisy street.
How we would use one
Look at the data weekly, not nightly. Focus on bedtime and wake-time consistency rather than the score. Use it to notice patterns worth mentioning to a clinician, then stop.
And if you find yourself checking it before you have got out of bed, or feeling defeated by a number before the day has begun, take it off at night for a fortnight. If you sleep better, you have learned something more valuable than anything in the app.
Frequently asked questions
Are sleep trackers accurate? Reasonably accurate for sleep timing and total sleep time, and considerably weaker for sleep stages, because they infer stages from movement and heart rate rather than measuring brain activity. They also tend to overestimate sleep in people who sleep badly, since lying still and awake resembles sleep to the sensors.
Can my smartwatch detect sleep apnoea? No. Some devices flag possible breathing disturbances or oxygen dips, which is a useful prompt to get tested, but none measures airflow and respiratory effort the way a diagnostic study does. A device that flags nothing does not rule the condition out.
Why does my tracker say I slept when I was awake? Because quiet, still wakefulness looks much like sleep to an accelerometer and a heart-rate sensor. This is a known limitation and it affects people with insomnia most, which is an unfortunate irony given they are the most likely to be checking.
Should I worry about my deep sleep score? Probably not. Stage estimates are the least reliable output these devices produce, and the score itself is a proprietary composite that changes between devices and firmware versions. How you feel during the day is a better guide than the number.
Is my sleep tracker making my insomnia worse? It can. Clinicians have described orthosomnia, where preoccupation with achieving ideal tracker data worsens sleep by turning bedtime into a performance evaluation. If you check your score anxiously or adjust your bedtime to improve it, try two weeks without wearing it.
This article is educational and is not a diagnosis or a substitute for medical advice. Slumbr does not provide emergency care. If you are in crisis, seek urgent in-person help.
Your tracker says one thing and you feel another? Take the free Slumbr Sleep Pattern Assessment™, which asks about your actual experience rather than your data. Or read how telephonic consultations work.