Sleep Score & Weekly Insights

The sleep score and its bands

The sleep score runs from 0 to 100 and sums up how well you slept that night. Your nightly summary shows the score with a label for its band:

If you wear a watch or ring that writes to Apple Health, the label also reflects your recovery score, so a night can read "Well Recovered" or "Short but Recovered" instead.

Duration carries the most weight. The app treats 7 hours or more as ideal, and that mark stays at 7 hours even if you set a different sleep goal. Falling short costs a little at first, since sleep need varies and a 6 to 7 hour night is fine for a lot of adults, and then the penalty gets steeper the shorter the night gets. After duration, the score looks at:

• Deep and REM percentages. The app targets roughly 13 to 23 percent Deep Sleep, with 20 percent or more earning credit, and about 20 to 25 percent REM. Landing in those ranges can add points. Estimating Deep and REM from audio is hard, so the app goes easy on low readings and counts quiet, restful Silence toward your restorative sleep. Sleep Stages explains why.
• Time to fall asleep. Taking longer than 30 minutes lowers the score.
• Continuity. How well you stayed asleep once you got there, built from your time awake after falling asleep (WASO) and your number of awakenings. The two combine into one capped penalty, so a fragmented night gets charged once instead of once per metric. Sleep efficiency still shows up as its own metric in your summaries, but inside the score, continuity has taken its place.
• Bedtime regularity. A bedtime within an hour of your usual one is neutral. Drift further than that and it costs points.
• Snoring and breathing disruptions. Snoring counts against the score based on how often it happens per hour. Breathing disruptions can pull it down noticeably, and the more there are, the bigger the hit.

Tap the score badge for a breakdown of how each factor affected your night. The score is a guide to your sleep habits. It isn't a clinical diagnosis.

A synced wearable can change the score too. When the app pulls sleep data from Apple Health, any night that also has a recording gets re-analyzed so the score can use the Deep and REM stages from your wearable, which are more accurate than the audio estimates. So a night's score may shift a bit once your Apple Watch finishes its staging and the data reaches the app. The wearable's recovery score feeds back into the sleep score as well. An unusually strong or weak night of vitals moves it up or down by as much as 15 points around a neutral baseline, and a typical night changes nothing. Apple Watch & Wearables covers the sync.

The recovery score

The sleep score rates how you slept. The recovery score is about how well your body recovered overnight. The sleep score reads your recording. The recovery score reads your overnight vitals, so it needs a wearable that writes to Apple Health, like an Apple Watch, Oura Ring, Whoop, Garmin, or Ultrahuman. Without one, the app shows your sleep score by itself and leaves recovery out.

Six inputs feed the score. Each one is measured against your own rolling baseline instead of a fixed target, so "good" means good for you:

• Heart rate variability (HRV). Higher variability points to a recovered nervous system. If your HRV is naturally low, you're never penalized for being below someone else's number, only your own baseline.
• Resting heart rate. The low your heart settled to overnight. A low rate, the kind fitness builds, lifts the score on its own.
• Respiratory rate. Steady, typical breathing scores well. A clear rise above your norm pulls it down.
• Blood oxygen (SpO2). Read against a healthy floor around 95 percent.
• Wrist temperature. Deviation from your baseline, on watches that record it.
• Time asleep. Scored against your sleep goal, 7 hours unless you change it, instead of your baseline, because sleep need doesn't bend much to habit.

The result is a 0 to 100 score, shown as a five-segment charge bar with a one-word verdict:

• Excellent: 82 or above, a strong night across the board
• Good: 74 to 81, clearly above your norm
• Fair: 58 to 73, at or near your baseline, the everyday default
• Low: 45 to 57, a real signal below your norm
• Poor: below 45

Holding your baseline lands you in Fair, and that's intentional. Maintaining is maintaining. You earn Good and Excellent by beating your norm. A short night caps the score no matter how good the vitals look, because you can't fully recover on too little sleep. Under 5 hours holds it to Low at most, and under 4 hours to Poor. Tap the score for the per-metric breakdown, where each vital shows its value, which way it moved against your baseline, and a plain-language note on what drove the night.

Since the recovery score already accounts for a short night, the sleep score's own duration penalty is kept moderate. A short night doesn't get charged twice across the two scores.

In the morning summary the two scores merge into one status line that pairs your time asleep with a recovery verdict, like "6h 42m · Well Recovered" or "Short but Recovered." The wording covers both how long you slept and how well you recovered. Like every metric here, the recovery score is for your own insight. It isn't a clinical readiness measure or a diagnosis.

Your morning summary

After each recording the app builds a nightly summary, so you can take in the whole night in one look. It shows your sleep score with its label, a breakdown of time in Light, Deep, REM, and Awake, your average respiratory rate in breaths per minute, and your snoring and breathing disruption activity. Tap the score badge and it expands into the factor-by-factor breakdown of what helped and what hurt.

The summary is the quick read on your habits and trends. The timeline is still the place to dig into individual sounds.

The Snoring row inside Sleep Insights does more than count. It compares your snoring rate in the minutes before each awakening against the night's average and says what it finds in plain words, for example that snoring climbed 66 percent just before your wakes, which hints that the snoring itself may be disturbing your sleep. To step through those wakes one at a time, open the snoring chart at the top of the same summary. That's covered in the next section.

The snoring chart

At the top of every nightly summary, under the headline count of snoring signals, the app plots the whole night as a bar chart. One picture shows how loud you snored, when you snored hardest, and what your sleep was doing at the time.

What the bars measure

Each bar covers a fixed slice of clock time. Short nights use 5-minute slices, stretching to 10 and then 15 minutes on longer nights so the bars never get too thin to read. Height is the average loudness of the snoring inside that slice, so a single cough can't set the height of a whole bar. Only snoring windows go into that average. The quiet stretches between them don't water it down.

Loudness runs on a scale from 32 to 72 dB, split into four even bands that set both the color and the height:

• Faint: under 40 dB
• Light: 40 to 48 dB
• Audible: 48 to 56 dB
• Heavy: 56 dB and above

These are the same bands the app uses to describe individual episodes, so a bar reading Heavy and an episode summary reading "Heavy snoring" mean the same thing. Episodes & Events covers that vocabulary. Under the chart, a key lists every band the night reached, with how long it lasted and what share of the snoring it made up. Snoring is drawn in saturated color and sleep stages in pastels, so you won't confuse the two scales.

Loudness and Rate

On nights with a lot of snoring, a Loudness and Rate switch sits beside the headline. Rate re-heights every bar to snores per minute and labels the busiest minute of the night, for example "Peak 15 snores/min · 2:40 AM". Color still means loudness in both modes, so switching changes what the height says and nothing else. Loudness is how bad it was. Rate is how relentless it was.

A readout sits on an actual bar instead of floating over the chart, and it names the figure that sums the night up. In Loudness mode that's the average, reading Avg 51 dB. In Rate it's the busiest minute. The readout also moves, stopping on a handful of bars and reading each one out with the time it started, so you can pull values off the chart without checking them against a legend. It pauses while audio is playing so the chart can follow the playhead.

The stage ribbon

When the night has sleep stages, from the app's own analysis or a synced wearable, a ribbon runs under the bars in the hypnogram colors, and the key beneath it gives each stage its duration and share of the night. Line the two up and you can see whether your loudest stretches sit in light sleep, and whether they land right before you woke.

Walking through your awakenings

Tap Awake in the stage key and the card turns into an awakening browser. A pager steps through every wake of the night one at a time, with its time and how long it lasted, like "Awakening 1 of 11 · 2:08 AM · 2m". The timeline above follows along, so each step lands you on that moment of the recording, ready to play.

Under the pager, the app describes that awakening in plain words. How far into the night it came, how long it ran, how hard you were snoring beforehand compared to your usual rate, which way the snoring was already heading, and how quickly it came back afterward. It sticks to what the timestamps can back up. You get "snoring came back at full strength 16 seconds later", and no claim that the snoring is what woke you.

Nudges on the chart

On nights when Snore Nudge fired, the key gets a third row counting the night's nudges. Tap it and a hairline marks each nudge on the chart, with its own pager and a verdict for every nudge. The snoring rate going in, what happened next, and whether you slept through it.

Sleep Efficiency

Sleep Efficiency is the share of your time in bed that you spent asleep. The app divides time asleep by total time in bed and reports the percentage. High efficiency means you fell asleep fast and stayed asleep with little time awake in bed. Low efficiency means the recording has a lot of wakefulness in it.

Snore Timeline rates the percentage on five levels:

• Excellent: 90% or above
• Optimal: 85 to 90%
• Good: 80 to 85%
• Fair: 75 to 80%
• Poor: below 75%

The Weekly Summary shows efficiency as a weekly average, which smooths out the odd restless night and gives you a steadier signal than any one recording.

WASO: time awake after falling asleep

WASO stands for Wake After Sleep Onset. It's the total time you spent awake after first falling asleep. Efficiency tells you how much of the night you slept. WASO tells you how broken up that sleep was once it started. Two nights with identical efficiency can feel very different if one packed its wakefulness into a single stretch and the other scattered it across the night.

The app rates WASO in five tiers:

• Minimal disruption: under 15 minutes
• Normal: 15 to 30 minutes
• Moderate disruption: 30 to 45 minutes
• High disruption: 45 to 60 minutes
• Severe fragmentation: 60 minutes or more

Lower WASO means more continuous sleep. Like every metric on this page, it's for your own insight and isn't a clinical measurement.

How Snore Timeline measures WASO

The WASO clock starts the moment the app first detects sleep. Wakefulness before that counts as your time to fall asleep, so a long wind-down in bed can't inflate your WASO. From then on, every awake period adds its length to the total.

The app reads wakefulness from movement sounds, room noise, and breathing that doesn't fit with sleep, and picking up your phone during the night counts as an awakening outright. A suspected awakening only counts after roughly 90 seconds with no sign of sleep, so the brief stirs everyone has between sleep cycles don't register. On nights where your sleep data comes from a synced wearable, WASO uses the wearable's awake staging instead of the audio estimate.

What counts as a normal WASO

Waking at night is a normal part of sleep. Adults surface briefly at the end of each sleep cycle, several times a night, and forget nearly all of it by morning. A WASO of 15 to 30 minutes sits in the app's Normal band, and under 15 minutes reads as minimal disruption. Sleep also gets lighter with age, so an older sleeper's typical WASO runs higher than a younger one's.

A single rough night says very little. Stress, a late meal, or an unfamiliar room can push one night into a higher band. The pattern across a week is the steadier signal, same as with efficiency.

WASO, awakenings, and efficiency

The app reports wakefulness three ways. Awakenings count how many times you woke. WASO totals how long you stayed awake. Sleep efficiency divides time asleep by time in bed, which folds your time to fall asleep into the picture. Ten one-minute wakes and one ten-minute wake give the same WASO with very different nights behind them, so the score breakdown checks your average wake length and words its verdict to match. Lots of brief wakes you settled back from, or a few long ones.

How WASO moves your sleep score

WASO feeds the continuity factor along with your awakening count. Up to 20 minutes awake costs nothing. Past that, the penalty steps up at 30, 45, and 60 minutes, waking more than a couple of times adds a little more, and the combined total is capped, so even a badly fragmented night can't crater the score through continuity alone. Tap the score badge and the Continuity row shows the exact points next to your night's numbers, for example "3 awakenings · 42m awake".

Lowering a high WASO

The recording will tell you more than the number does. Open the timeline around an awake period and listen to what came right before it. A loud snore, a noise in the room, a partner, a pet. Waking after snoring or a breathing disruption points somewhere different than waking after a barking dog.

Then test your suspects with sleep factors. Alcohol tends to fragment the second half of the night, and Late Meal, Caffeine, Stress, Hot Room, and Nap are common culprits worth tagging. After enough tagged and untagged nights, the comparison shows whether your WASO differs in your own data, and the Test a Remedy guide walks through it. A steady bedtime helps too. The Consistency Score tracks that.

If a high WASO shows up alongside frequent breathing disruptions, that combination is worth showing a doctor. Export & Sharing turns your nights into a report you can bring along, and the CSV export has a WASO column for every night.

Bedtime Consistency Score

Going to bed at a regular hour matters enough that the app scores it separately. The Consistency Score reflects how regular your bedtime has been over the last 14 days, based on how much it varies night to night. There's a 1-hour buffer. Variation up to about an hour still scores 100, and the score drops as your bedtimes spread out beyond that.

The app takes each night's bedtime from your earliest recording start time, so starting the recording when you get into bed keeps this metric accurate. The Weekly Summary also shows your average Bedtime for the week.

Your sleep goal

Snore Timeline measures your nights against a sleep goal of 7 hours until you pick your own. To change it, open Settings, find Sleep Goal under Detection, right below the Sleep Insights toggle, and choose anything from 6 to 10 hours in 30-minute steps.

The goal sets the target for your Sleep Bank, the Sleep Deficit below, your sleep streaks, the goal bar on the Live Activity and widgets while you record, and the time-asleep row of the recovery score. Your Apple Watch picks up the new goal too. Nights you've already recorded keep the score they earned, and the 0 to 100 sleep score keeps its own 7-hour mark for duration.

Sleep Deficit

Sleep Deficit lives in the Weekly Summary and shows the gap between how much you slept and your nightly sleep goal, added up across the whole week. A positive number means you fell short over the week. A negative number means you slept more than the target, a surplus.

It sits next to Time Asleep, which is your average nightly duration for the week. Time Asleep tells you what a typical night looked like. Sleep Deficit tells you what the whole week added up to. A few short weeknights can hide inside a decent average and still leave a deficit you can see right away.

Sleep factors and correlations

Sleep factors turn your recordings into experiments. You tag what you did, and the app compares tagged nights against untagged ones, so the patterns come out of your own data instead of guesswork.

What you can track

The app includes 23 built-in factors across four categories:

• Substances: Alcohol, Caffeine, Medication, Smoking
• Lifestyle: Exercise, Late Meal, Stress, Late Screen Time, Nap, Fatigue
• Environment: Slept on Back, Slept on Side, Elevated Pillow, Hot Room, Cold Room
• Health: Nasal Strips, Congestion, Allergies, Cold/Flu, CPAP, Mouth Guard, Oral Appliance, Mouth Taping

If the list is missing something, create a custom factor by giving it a name and picking one of those four categories. You tag factors on your nightly summaries with Tag Sleep Factors (it reads Edit Sleep Factors once a night has some), and swipe the picker down when you're done. A factor you have every night, like CPAP or a mouth guard, can be set to Repeat nightly so it tags itself. The trick is still to tag selectively. Only tag a factor on nights you had it. Tag something every single night and there are no factor-free nights left to compare against, so no correlation can show up.

Building Insights and unlocking comparisons

Correlations need data before they mean anything. The weekly comparison needs at least 5 recorded nights in total. Nights with synced wearable sleep data count too, even without an audio recording. Until then, a Building Insights state shows a progress bar, how many more nights you need, and the factors you've tracked so far. After that, a factor gets its row as soon as you tag it, but the app only states an effect once the factor has at least 3 recorded nights with the tag and 3 without. Before that, the row tells you how many more to tag. The app looks back over your last 8 weeks of data to build the patterns, and more nights make the results more confident.

Reading the results

Once unlocked, the "What Affected Your Sleep" card in the Weekly Summary ranks your factors as plain-language statements, strongest first, for example "Elevated Pillow cut breathing disruptions 72%". Each row carries a verdict based on an impact score. Better sleep means the factor is associated with improvement in your data, Worse sleep with a decline, Mixed signals means some metrics moved your way while others didn't, and No clear effect means the difference between your with and without nights was too small to call.

Tap a row to expand it. The expanded view shows how much each metric changed on tagged nights versus untagged ones, with a change bar per metric. Snore time, events per hour, breathing disruptions, sleep talking, sleep score, and sleep efficiency, plus resting heart rate, HRV, and breathing rate when a wearable feeds Apple Health. Those wrist vitals go into the impact score too, so a factor that mostly moves your physiology, like stress, can still come out as your strongest. Below the bars there's a short line on the likely mechanism, a coaching tip, and a footer with the night counts for each group and the reminder "Association, not cause." Factors that haven't crossed the thresholds yet show how many more tagged nights they need instead of expanding.

Take CPAP, which is a built-in factor. Tag it on the nights you used your machine and leave the other nights untagged. Once you have 3 recorded nights in each group, the CPAP row shows whether your machine nights had less snore time, fewer disruptions, and a higher score. The Breathing Disruptions page and the Test a Remedy guide walk through this in full.

A factor up close

Tap any factor chip on a nightly summary to open its detail sheet. It has four parts:

• Your Result. How the factor lines up with your sleep, once you have 3 recorded nights with it and 3 without. Until then, three dots fill in as the nights come in.
• Getting It Right. For every built-in factor: what works, the common mistakes, how long to give it, and when to see a professional. Changes you're testing come with a trial length, for example 14 nights for side sleeping or a mouth guard, 7 for nasal strips or an elevated pillow, and 30 for CPAP, and the sheet counts your nights against it. Custom factors don't get a guide.
• Your Nights. How many recorded nights in the last 8 weeks had the factor and how many didn't, when you last tagged it, and a 4-week strip of the nights you tagged.
• How It Affects Sleep. A short note on the likely mechanism.

The guidance is general, not medical advice, and the sheet says so.

Correlations show patterns in your own recordings, and they may reflect other habits that happened at the same time. Treat them as personal insight to discuss with your doctor. They aren't cause and effect, and they aren't a diagnosis.

Weekly trends

The Weekly Summary gives you a 7-day overview of your snoring and sleep data and compares it against the previous week. Each metric shows a week-over-week change, as a percentage for most of them. Peak Volume, measured in decibels, shows the difference in dB instead. When the previous week had no data, the change reads 0% in a neutral color, so you don't get a trend flagged that doesn't exist. Daily charts plot each metric across the last 7 days so you can see the shape of the week as well as its total.

The summary averages a broad set of metrics:

• Snoring: Snore Time, Snoring Signals, Signals Per Hour, plus breathing disruptions and sleep talking (Sleep Talking Time and Talking Signals).
• Sleep quality: Sleep Score, Time Asleep, Sleep Efficiency, Sleep Deficit, Deep Sleep (ideal 13 to 23%), REM Sleep (ideal 20 to 25%), Awakenings, Bedtime, and the Consistency Score.
• Highlights: the Top 5 Notable Episodes from the week, which link back to episode details.

Your average Sleep Score for the week has its own week-over-week change and 7-day chart, so you can watch the score move night by night instead of judging one morning.

Most people open this screen to find out whether their snoring is getting better or worse. For that, watch three numbers. Snore Time is total time spent snoring. Snoring Signals is the total count of snore-related events, including snoring, coughs, gasps, and disruptions. Signals Per Hour is signals per hour of recording. If those trend down from one week to the next, you're snoring less overall, and the daily charts show which nights drove the change.

These are personal metrics for tracking patterns over time. To share them with a doctor, Export & Sharing shows how to turn a week of data into a report.