Sleep: the most important things, ranked

Most sleep advice is a flat list of twenty tips, which hides the fact that the items differ in effect size by an order of magnitude. Here they are in order, with the reasoning for each placement.

Most sleep advice is a flat list of twenty tips, which is useless, because the items differ in effect size by an order of magnitude. Blackout curtains and undiagnosed sleep apnea do not belong in the same sentence. Here they are ranked, with the reasoning.

1. A consistent wake time

Not bedtime. Wake time.

Your circadian clock is set primarily by when light first hits your eyes, so a fixed wake time anchors everything downstream: when you get sleepy at night, when your energy peaks, how deep your sleep goes. A steady wake time does more than a perfect bedtime ever will, because you cannot force sleep onset but you can control when you get up.

The evidence for regularity has moved recently. A 2024 UK Biobank analysis of accelerometer data from around 60,000 people found sleep regularity predicted all-cause mortality better than sleep duration did. That is a striking result: how consistent you are mattered more than how long you slept.

This is first because it is free, it works for nearly everyone, and it makes every other item on this list easier.

2. Getting screened if you snore or wake up unrefreshed

This one is different in kind from everything else here, which is why it sits so high.

Obstructive sleep apnea is a medical condition, not a habit. It fragments sleep hundreds of times a night without waking you enough to remember, and it is strongly associated with depression, hypertension and cardiovascular disease. Estimates suggest the large majority of moderate to severe cases are undiagnosed.

No amount of sleep hygiene fixes it. If you snore heavily, wake with a dry mouth or headache, or feel unrefreshed after eight hours, the useful next step is a screen and then a proper sleep study, not a better pillow. Mura's sleep screen listens for breathing pauses and snoring overnight as a first pass, and why sleep hits mood so hard covers the mechanism. Either way, a positive result means see a doctor, not self-treat: a screen is not a sleep study.

3. Light, in the right direction at the right time

Bright light within an hour of waking, ideally outdoors, since even an overcast morning beats indoor lighting by a wide margin. Then dim light in the last hour or two before bed.

Light is the strongest input your clock has. This is the mechanism that makes item 1 work, which is why it sits immediately below it.

And darkness once you are asleep. The HEIJO-KYO cohort, run by Kenji Obayashi's group at Nara Medical University, measured bedroom light in the homes of several hundred older Japanese adults rather than asking them to estimate it. People sleeping with 5 lux or more had roughly double the rate of developing depressive symptoms over the following couple of years compared with those sleeping in near darkness. The same cohort has linked night-time light to obesity, raised blood pressure and poorer glucose control.

Two things worth holding alongside that. It is observational, in an older cohort, so reverse causation is live: early depression changes sleep habits too. And 5 lux is a useful calibration rather than a scare figure. It is roughly a dim nightlight, or a streetlight through thin curtains. It is not a standby LED on a television, which sits far below that. Blackout curtains or a sleep mask are worth the small effort. Taping over every small indicator light is not the thing that decides this.

4. If you have chronic insomnia: CBT-I, ahead of everything else here

If you have been sleeping badly for months rather than days, this list is the wrong list. Cognitive behavioral therapy for insomnia is first-line treatment in both American Academy of Sleep Medicine and American College of Physicians guidelines, ahead of medication, and its benefits hold after treatment ends in a way sleeping pills' do not.

The counterintuitive core of it is spending less time in bed, not more, so that bed re-associates with sleep rather than with lying awake. That is the opposite of what most people do, and it is why the self-help version usually fails without structure.

5. Alcohol

The most consistently underestimated item on the list.

Alcohol shortens sleep onset, which is why people believe it helps, then suppresses REM early in the night and fragments the second half. You spend hours in bed and get very little of the emotional processing sleep is for. It also raises heart rate and suppresses HRV overnight, so you wake with your nervous system already braced. We went into that mechanism in hangxiety.

If you drink and sleep badly, moving this up your own list is usually the fastest measurable change available.

6. Caffeine, and specifically its timing

Caffeine's half-life is roughly 5 to 6 hours, so a 4pm coffee still has meaningful drug in your system at midnight. People who insist caffeine does not affect them are often right about sleep onset and wrong about sleep depth, which they cannot feel.

A hard cutoff around eight hours before bed costs nothing and is easy to test on yourself.

7. A cool room

Core body temperature has to fall for sleep to begin. Around 18°C (65°F) is the usual recommendation. Real, worth doing, and far less important than anything above it.

8. Exercise, with a timing caveat

Regular activity improves sleep quality, and the effect is reliable if modest next to items 1 through 5. Vigorous training very close to bedtime delays onset for some people and not others, so this is worth testing rather than obeying.

What did not make the list, and why

A ranking is only useful if it also tells you what it left out. These are the things most sleep advice leads with, and the reasons they are not here.

Melatonin. It is a circadian signal, not a sedative. It tells your body what time it is; it does not make you sleepy in the way people expect. That makes it genuinely useful for jet lag and for delayed sleep phase, where the problem is that your clock is in the wrong place. For ordinary "I cannot fall asleep", the 2017 American Academy of Sleep Medicine guideline recommends against using it for sleep onset or sleep maintenance insomnia in adults. That is graded a weak recommendation rather than a prohibition, but weak in the direction of not bothering. Worth knowing too that the doses sold over the counter are often many times what your body would ever produce, and more is not better for a timing signal.

Magnesium. The evidence is thin: small trials, often in older adults, often in people who were deficient to begin with. If you are deficient, correcting that is worth doing for reasons well beyond sleep. As a sleep intervention for everyone else, it is nowhere near items 1 through 5.

Blue-light blocking glasses. A 2023 Cochrane review of blue-light filtering lenses found no clear benefit for sleep quality or eye strain. The underlying biology of light and the body clock is real, which is item 3, but the glasses specifically have not delivered.

Sleep trackers. These have a failure mode with its own name, orthosomnia: anxiety about the score itself starts damaging the sleep it is measuring. If your tracker tells you that you slept badly and you then spend the day braced for it, the tracker has cost you more than it gave. Useful for spotting a pattern over months. Not useful checked every morning.

Eight hours as a target. Sleep need varies between people, and chasing a number turns sleep into a performance with a pass mark. Effort is the one thing that reliably prevents sleep onset, which is why the harder you try the worse it goes.

How long any of this takes

Worth setting expectations, because the usual failure is trying everything for three nights and concluding that nothing works.

Removing alcohol shows up the very next night. It is the fastest measurable change on this list, which is part of why it ranks where it does.

A consistent wake time and morning light need one to two weeks. You are moving a biological clock, and it shifts by roughly an hour a day at best. The first few mornings feel worse, not better, because you are getting up at a fixed time while still falling asleep at your old one.

CBT-I is typically four to eight sessions across a similar number of weeks, and it usually feels harder before it feels easier, since the core of it is spending less time in bed.

Caffeine timing and room temperature are immediate but small. They are worth doing and they will not rescue a schedule that is fundamentally irregular.

The short version

Fix your wake time. Rule out apnea if you snore. Get light right. If it has been months, get CBT-I. Then, and only then, worry about the room temperature.

This is general information, not medical advice. Persistent sleep problems, loud snoring or witnessed breathing pauses should be assessed by a doctor.

This article is for general information and is not a substitute for professional care. If you are in crisis or thinking about harming yourself, please contact your local emergency services or a crisis line. See crisis lines.