What is a good sleep score? What the number means and how to improve it

Short Answer

A sleep score is a single 0–100 number your wearable gives you each morning, bundling duration, efficiency, timing, stages and continuity into one figure. Most devices treat 80–89 as good and 90–100 as excellent. The formula is proprietary, so scores are not comparable between brands.

Three caveats matter more than the number itself. A sleep score is a single 0–100 number your wearable gives you each morning that bundles several aspects of the night — how long you slept, how efficiently, when, your mix of sleep stages, and how continuous it was — into one easy figure. As a rough guide, most devices treat 80–89 as "good" and 90–100 as "excellent," with lower scores flagging shorter, more broken, or less well-timed sleep. But three big caveats matter more than the number itself. First, the exact formula is proprietary and secret, and every brand weights things differently, so your score isn't comparable to a friend's on another device. Second, the stage estimates it's built on are approximations, not lab-grade measurements. Third — and most important — how you actually feel matters more than the score, and obsessing over the number can itself harm your sleep. So use your sleep score as a rough, personal trend indicator and a nudge toward better habits, not as a grade to chase. This article explains what goes into the score, what counts as good, the caveats, and how to genuinely improve it.

A note on the data: Welltory looks at your sleep and recovery signals, so it can show how your sleep quality trends over time and what moves it. The useful thing about any sleep score is not the single morning's figure but the pattern — whether your sleep is generally solid and how it responds to alcohol, late nights, stress, and schedule. Because the underlying stage estimates are approximate and every brand scores differently, the score is best read as a personal trend and a prompt, not an exact grade.

What is a sleep score?

Modern wearables — watches and rings — track signals through the night (movement, heart rate, sometimes blood oxygen and temperature) and use them to estimate your sleep. Rather than make you interpret a dozen metrics, they roll everything into one number from 0 to 100 so you can glance at your night in a single moment. (biologyinsights.com) It's a genuinely convenient summary — a quick "how was my sleep?" — and for many people it's the single most-looked-at health number they own, checked before they've even gotten out of bed. The key is understanding that it's a composite estimate: a blend of several things the device measured, combined by a formula the manufacturer designed. That makes it useful for spotting patterns, but also means it's only as good as its inputs and its recipe — both of which have real limits.

What goes into a sleep score

While the exact recipe is secret, most sleep scores combine the same broad ingredients. (sleepspace.com)

Duration. How long you actually slept — the biggest factor for most scores, with most adults needing 7–9 hours.

Sleep efficiency. The percentage of time in bed that you spent asleep; around 85% or higher is considered good, meaning you weren't lying awake much.

Sleep stages. The balance of your night, especially how much deep and REM sleep you got — the restorative stages. (See how much deep sleep you need and how much REM you need.)

Continuity. How unbroken your sleep was — lots of awakenings or restlessness pull the score down.

Timing and consistency. When you slept relative to your body clock and your usual schedule; erratic or badly-timed sleep scores worse.

Heart rate and HRV. Many devices fold in your overnight resting heart rate and HRV as markers of how well you recovered. (See HRV during sleep.)

The score is essentially these signals, weighted and blended. That's why a night that was long but restless, or full but badly timed, can score lower than you'd expect — the composite is trying to capture quality, not just hours — which is genuinely more informative than duration alone, even if it's harder to interpret at a glance.

A closer look at sleep efficiency

Of the ingredients, sleep efficiency is worth understanding on its own, because it's often what quietly drags a score down. Efficiency is simply the percentage of your time in bed that you actually spent asleep — so if you were in bed eight hours but only slept seven, your efficiency is about 88%. Around 85% or higher is considered good; much lower means you're spending a lot of time lying awake, whether struggling to fall asleep, waking in the night, or drifting before your alarm. This is why "I was in bed for nine hours" doesn't guarantee a great score: if two of those hours were spent awake, your efficiency — and your score — takes a hit. It's also a useful diagnostic in itself. Low efficiency from trouble falling asleep points to wind-down and stimulation issues (screens, caffeine, a racing mind); low efficiency from waking during the night points to things like alcohol, temperature, or fragmented sleep; and consistently poor efficiency despite good habits can be a reason to talk to a doctor. Improving efficiency — falling asleep faster and staying asleep — is one of the most direct ways to lift a sleep score.

What is a good sleep score?

On the usual 0–100 scale, the common bands are: 90–100 "excellent," 80–89 "good," and progressively lower scores indicating more room for improvement, with scores in the 70s often labeled "fair" and below that "poor." (simplewearablereport.com) So if you're regularly in the 80s, your sleep is generally solid; consistent 90s are excellent; and frequent scores in the 60s–70s suggest your sleep could use attention. But hold these bands loosely, because — as we'll see — the number is a rough estimate and the exact thresholds vary by brand. The more meaningful question than "is 84 good?" is whether your score is generally in a healthy range for you and stable or improving, and whether you wake feeling rested. A single lower night is normal noise; a persistent downward trend, or a good-looking score paired with feeling awful, is what's worth attention.

The caveats that matter more than the number

Here's where honesty is essential, because sleep scores are easy to over-trust.

The formula is a black box. Manufacturers don't disclose their exact algorithms, component weights, or thresholds, so you can't fully know why you got a given score. (sleepspace.com)

Scores aren't comparable across brands. Because each company uses a different recipe, an 85 on one device is not the same as an 85 on another. Comparing your score to a friend's on a different watch is meaningless.

It's built on estimates. The stage and sleep detection underneath are approximations from movement and heart rate, not the brain-wave measurement of a lab study, so the components — and therefore the score — can be off on any night.

Chasing the number can backfire. There's a recognized phenomenon, sometimes called orthosomnia, where fixating on perfecting a sleep score creates anxiety that actually worsens sleep. The score is meant to inform, not to become a source of bedtime stress.

None of this makes the score useless — it's a handy trend indicator and habit nudge. But it does mean you should treat it as a rough guide, cross-checked against how you feel, rather than a precise verdict.

How do I improve my sleep score?

The good news is that improving your score isn't about gaming an algorithm — it's about genuinely sleeping better, and the score largely follows. The highest-impact moves are the familiar fundamentals: get enough total sleep (aim for your 7–9 hours; duration is usually the biggest lever); keep a consistent sleep and wake schedule, which helps timing, continuity, and stages all at once; avoid alcohol near bedtime (it hurts efficiency and suppresses deep and REM sleep); skip late heavy meals and late caffeine; keep the room cool, dark, and quiet; get daylight and exercise during the day; and wind down before bed to fall asleep faster and improve efficiency. Do these consistently and your duration, efficiency, stage balance, and continuity all tend to improve — and the score with them. What doesn't help is fixating on the number itself. Build good sleep, and let the score be a byproduct. It's worth adding that the single biggest lever is almost always duration — simply giving yourself enough time asleep. Many disappointing scores come down to not spending enough hours in bed rather than to some subtle quality problem, so before optimizing anything exotic, check that you're actually allowing yourself a full 7–9 hours. Fix the hours first, and a surprising number of "poor sleep" complaints resolve on their own, with the score rising to match. (These overlap almost entirely with the habits in how to improve your HRV, since good sleep drives both.)

How long before you know whether the levers were the answer. The levers move at two very different speeds, and treating them as one is why people quit after four days. The fast ones show up within a week: give yourself enough time in bed, stop drinking in the few hours before sleep, move heavy meals and late caffeine earlier — you are removing something rather than retraining something, and duration in particular tends to lift a score almost immediately. The slow ones need two to three weeks: a consistent bed and wake time has to re-anchor your body clock before timing, continuity and stage balance settle, and a genuinely lower stress load takes at least that long to reach your nights. So the honest test is one week on hours in bed and evening alcohol, then two to three weeks on a fixed schedule and a real wind-down — and judge it by how rested you feel on waking, with the score trend as a second opinion rather than the verdict. If you get to the end of three honest weeks and the mornings are still bad, stop optimising the score. That is the informative result, not a failure, and three documented weeks is exactly what makes the conversation below land.

Why does the same night score differently on two apps?

A subtle but useful thing to understand: because sleep scores weight their ingredients differently, the same night can produce different scores and even different verdicts on different devices — one might penalize a couple of awakenings heavily while another cares more about total duration. This isn't a malfunction; it's a design choice, and it's a reminder that a sleep score is an opinion about your night, formed by a particular algorithm, rather than an objective truth. It also explains why people who switch devices sometimes see their "typical" score jump or drop overnight even though their sleep hasn't changed — they've simply moved to a different scoring philosophy. The practical takeaway is to anchor on one device and its trend, and not to read too much into the absolute value or into changes that coincide with switching hardware. Your sleep didn't get better or worse because you changed watches; only the opinion about it did.

When a low score is worth acting on

Most single low scores are just noise — a late night, a glass of wine, a stressful day, or a restless night that everyone has. What deserves attention is a persistent pattern: your score sitting low for weeks, especially if it lines up with genuinely feeling unrested, tired, or foggy. That combination — a sustained low trend and poor subjective sleep — is the signal to act, first by tightening the fundamentals (duration, consistency, alcohol, wind-down) and, if that doesn't help, by talking to a doctor, since persistently poor sleep can have treatable causes like insomnia or sleep apnea. On the other hand, a low score on a morning you feel fine rarely warrants any action at all. The skill is distinguishing the isolated dip (ignore) from the entrenched trend (act), and letting how you feel weight the decision. A score is most valuable not for grading last night, but for flagging when something has shifted over weeks that's worth looking into.

Why does my sleep score disagree with how I feel?

The most important calibration: your subjective experience is the ultimate measure of sleep, and the score is a helpful proxy, not a replacement for it. If you wake up genuinely rested, energetic, and clear-headed, your sleep was good regardless of what the number says — a "disappointing" 76 on a morning you feel great is not a problem. Conversely, if you feel exhausted and foggy despite a high score for days on end, that mismatch is itself worth attention, and it's a reason to look further (or see a doctor) rather than trust the number over your body. The healthiest way to use a sleep score is as one input among several — alongside how you feel, your energy, your mood, and your other recovery signals — with your lived experience as the tiebreaker. Used that way, it's informative and motivating; treated as the final word, it can mislead you and stress you out.

How to bring this up with your doctor — and what to ask for

"My sleep score is low" is the version of this complaint most likely to be waved away, and not unfairly — no clinician can act on a proprietary number from a wrist. What they can act on is unrefreshing sleep that has lasted for months and is costing you your days. Lead with that, and keep the score as the second page.

Bring two weeks of a simple log rather than a screenshot of your worst score. Bedtime, wake time, how rested you felt in the morning, and one word for what the evening contained — alcohol, late meal, hard day, workout, nothing unusual. A pattern someone else can read carries weight that a description does not, and it is the only form in which tracker data helps inside a ten-minute appointment. The most useful line you can add is the mismatch itself, if you have one: "the nights the app calls good feel the same as the nights it calls bad."

Expect a lab form and a referral, not an answer. Apnea is settled by a sleep study, thyroid and iron by blood work, so the realistic best outcome of one visit is the next test rather than a diagnosis. If you leave with those, the visit worked — and knowing that in advance is what stops a referral from reading as a brush-off.

How Welltory helps

Welltory looks at your sleep and recovery signals and shows them as a trend over time, which is exactly how a sleep score is best used — not as a single morning's grade but as a pattern you can learn from. You can see whether your sleep is generally solid and, more usefully, connect it to what you did: the dip after drinking or a late night, the improvement across a consistent stretch, the effect of stress. That turns a vague number into actionable insight about what actually helps your sleep. You can also view your sleep alongside your overnight heart rate and HRV for a fuller recovery picture than any single score gives. Two honest caveats: sleep-stage estimates are approximate and scoring differs by device, so read your own trend rather than an exact figure or a cross-brand comparison; and let how you feel be the final judge — a number should never override waking up rested or exhausted. But as a way to see what's helping and harming your sleep over time, it's genuinely useful. (See also our guide to sleep trackers and why 8 hours can still leave you tired.)

What your data cannot see, stated plainly. Welltory does not measure brain waves, so sleep and its stages are estimated rather than measured — and any score built on estimates inherits their error. It does not measure your breathing in a way that could say anything about apnea, and it does not diagnose insomnia, restless legs, a thyroid problem, anemia or depression; a wearable cannot rule any of them out either. It cannot tell you why a night went badly — only that it looked that way, and what the evening around it contained. It does not know how much sleep you personally need, and it cannot see the things that broke your night from outside you: a child, a noisy street, a partner, pain. And it records trends, not verdicts: one night is noise, and a figure from another brand is not comparable to yours. What it does record is sleep timing, heart rate, heart rate variability and stress load, which is enough to show which of your own inputs move your nights. If you feel exhausted for weeks while the numbers look fine, believe your body and take that to a doctor — that mismatch is information, not a tracking problem.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data