Sleep Trackers: Which Metrics to Ignore (and Which Matter)

A 2021 head-to-head test of four popular wearables against laboratory polysomnography found the devices detected sleep with 91 to 97 percent sensitivity — but correctly identified wakefulness only 19 to 52 percent of the time. In plain terms: when you lie still scrolling your phone at 2 a.m., your tracker probably logs it as sleep.

That matters because millions of first-time users now steer their evenings by these numbers — shifting bedtime, cutting caffeine, ordering supplements — on data that is partly measurement noise. Knowing which metrics to ignore is not a technicality. It is the difference between a useful habit and a new source of anxiety.

What Consumer Trackers Can and Cannot Measure

About one in three US adults sleeps less than the seven hours per night recommended by the CDC, and the NHS puts the healthy range for most adults at six to nine hours. Sleep tracking is now a default feature on mainstream watches and rings, so the audience for these numbers is enormous — and largely untrained in reading them.

Consumer devices estimate sleep from movement (actigraphy), heart rate and heart-rate variability. They do not record brain activity. Polysomnography, the lab gold standard, measures EEG brain waves, eye movements and muscle tone. Because sleep stages are defined by brain-wave patterns, a wrist device is always inferring them from a pulse.

The American Academy of Sleep Medicine said this explicitly in a 2018 position statement: consumer sleep technologies are not validated for diagnosing sleep disorders and should not drive clinical decisions. Validation studies also show wearables tend to overestimate total sleep time — by 30 to 60 minutes in some studies, especially among people with insomnia — because stillness gets scored as sleep.

Why It Matters

Beginners tend to treat every metric on the dashboard as equally trustworthy. They are not, and the gap is practical, not academic.

Sleep researchers have already named the failure mode. In a 2017 case series in the Journal of Clinical Sleep Medicine, investigators coined the term "orthosomnia" for patients so preoccupied with achieving perfect tracker data that they resisted effective treatment or slept worse — in some cases because the device insisted sleep was fine while lab testing showed a treatable disorder.

The wrong metrics push you toward three dead ends: chasing deep-sleep percentages you cannot directly control, worrying over oxygen dips that carry no diagnostic value, and treating normal night wakings as failures. Adults typically surface briefly several times per night; a tracker converts that normal physiology into a red graph.

The Metrics You Should Ignore

Sleep-stage percentages

Deep, light and REM breakdowns from a wrist or finger are educated guesses. Without EEG, agreement with lab staging is modest, and two devices worn on the same wrist can disagree substantially. Read the chart as a rough story, never as a measurement.

Proprietary "sleep scores"

Every brand runs its own formula, and none is published or clinically validated. A 78 on one app is not a 78 on another. If a number changes your mood more than your behavior, it is working against you.

Nightly heart-rate variability

Single-night HRV swings with alcohol, late meals, hard training, illness and stress. Only multi-week trends carry any signal, and even then HRV is context, not a verdict on last night.

Blood-oxygen and breathing-disturbance readings

Consumer SpO2 sensors are not medical devices. They can produce false alarms and false reassurance, and they cannot diagnose sleep apnea. If you snore heavily, wake gasping or feel unrefreshed despite adequate time in bed, ask a clinician about a home sleep apnea test.

Minute-by-minute "awake" flags

Brief awakenings are part of normal sleep architecture, not defects. A dashboard that highlights each one makes healthy sleep look broken.

What to Watch Instead

• Bedtime and wake-time consistency across the week — the most actionable signal a tracker gives you.
• Weekly average sleep duration, checked against the seven-to-nine-hour range most adults need.
• Time in bed versus estimated sleep time, as a crude efficiency check: nine hours in bed for six and a half hours of sleep is a solvable scheduling problem.

Regularity has genuine evidence behind it. A 2024 UK Biobank analysis published in the journal Sleep found that sleep regularity predicted mortality risk more strongly than average sleep duration. Duration and consistency are also the two variables you can change tonight, no algorithm required.

The Honest Counterpoint

Sleep researchers do defend consumer trackers, and the defense is not trivial. Actigraphy is a legitimate method used in sleep research for decades. Modern devices perform better than early models, tracking raises awareness that helps some people build steadier routines, and wearable trends can complement the sleep diaries used in cognitive behavioral therapy for insomnia.

That objection is fair — and it still does not rescue the granular metrics. The validated use case is sleep-versus-wake timing and trends over weeks, not stage percentages, nightly scores or oxygen graphs. The same heightened awareness that helps one user feeds orthosomnia in the next.

Our editorial take, labeled as opinion: treat the tracker as a logbook, not a laboratory. Use it to check whether your bedtime is drifting and whether your weekly average sits near seven hours. Ignore everything the app presents as a grade. If checking the data makes you anxious, stop looking at it nightly — that is not quitting, it is correct use of the tool.

Frequently Asked Questions

Are sleep trackers accurate for sleep stages?

No. Stages are inferred from movement and heart rate rather than brain activity, and validation studies show only modest agreement with polysomnography.

What is orthosomnia?

A term coined in a 2017 case series for perfectionistic preoccupation with achieving perfect tracker sleep data — a pattern that can itself worsen sleep.

Can a smartwatch detect sleep apnea?

No. Consumer oxygen and breathing metrics are not diagnostic. Ask a clinician about a home sleep apnea test if you have symptoms.