Sleep is not a switch you flip. It is a hill you coast down, and a wind-down routine is just building yourself a gentler slope.
The useful part is that the three things most worth changing have all been measured properly, and the numbers are larger than the advice usually implies.
Key takeaways
Ordinary room light before bed suppressed melatonin and shortened its duration by about 90 minutes, in 99% of people tested (Gooley et al. 2011, JCEM).
Reading on a light-emitting screen before bed delayed the body clock, suppressed melatonin, and reduced next-morning alertness compared with a printed book (Chang et al. 2015, PNAS).
Caffeine 6 hours before bed measurably disrupted sleep, which is why the standard advice is a six-hour cutoff (Drake et al. 2013, Journal of Clinical Sleep Medicine).
A repeated routine becomes a cue, which is a second benefit on top of the direct effects.
Why does the descent matter?
Because your brain needs a runway between doing things and being unconscious.
Jump straight from bright rooms and busy thoughts into bed and you will often lie there while your mind finishes the day without your permission. A repeated wind-down works two ways: it lowers arousal directly, and, done consistently, the routine itself becomes a signal. After a few weeks the first step starts making you sleepy, the way a familiar driveway makes you feel home.
That second mechanism is why consistency beats elaborateness. A boring routine you repeat is worth more than a luxurious one you do occasionally, which is the same lesson sleep consistency keeps teaching.
How much does ordinary light matter?
Far more than most people assume, and this is the most underrated item on the list.
Researchers compared normal room light with dim light in the hours before bed. Compared with dim light, room light suppressed melatonin and produced a later melatonin onset in 99.0% of individuals, shortening the body's internal representation of night by about 90 minutes (Gooley et al. 2011).
Room light during the usual hours of sleep suppressed melatonin by more than 50% in 85% of trials.
Note what is being described. Not a screen, not a bright lamp in your face. Ordinary domestic overhead lighting, the kind most people sit under until they go to bed. Ninety minutes is a substantial shift in your body's sense of when night begins.
The fix is unusually cheap: lamps instead of overheads for the last hour. You do not need to sit in darkness, just lower.
Are screens really the problem?
Yes, and for two reasons rather than one.
In a controlled comparison, participants reading an electronic book on a light-emitting device before bed took longer to fall asleep, felt less sleepy in the evening, secreted less melatonin, had a later circadian clock, and were less alert the next morning than when reading a printed book (Chang et al. 2015).
That is the light half. The other half is content. A device combines short-wavelength light with material engineered to keep you engaged, and the feed is usually the bigger problem than the glow. A dimmed screen showing something absorbing is still a device keeping you awake.
The practical move handles both at once: put the phone outside arm's reach. Not off, not filtered, just physically elsewhere. It solves the light and the pull in one action.
How late is too late for coffee?
Six hours is the evidence-backed answer, and it surprises people.
A study gave participants 400 mg of caffeine at 0, 3 and 6 hours before their habitual bedtime, against placebo, and measured sleep both by self-report and with a validated portable monitor. All three timings significantly disrupted sleep compared with placebo (Drake et al. 2013).
The authors' conclusion was explicit: the reduction in total sleep time from caffeine taken six hours before bed was large enough to support the standard recommendation to avoid substantial caffeine for at least six hours before bedtime.
Two honest caveats. 400 mg is a decent dose, roughly four cups of brewed coffee. And caffeine sensitivity varies genetically by a lot. But the direction is clear, and the six-hour figure has actual experimental support rather than being folklore. An early-afternoon cutoff is kind to your midnight self.
Worth noting for a different metric: caffeine reliably raises blood pressure in non-habitual drinkers but does not reliably raise heart rate, so a racing pulse is not the mechanism by which it keeps you awake.
A thirty-minute blueprint
Keep it boring and repeatable. Three ten-minute movements is plenty.
Minutes 0 to 10: lower the lights. Overheads off, lamps on. This is the highest-value ten minutes in the routine given the 90-minute melatonin effect.
Minutes 10 to 20: park the phone. Physically out of reach, in another room if you can. Handles light and content together.
Minutes 20 to 30: something dull and analogue. A printed book, tidying, a shower, slow breathing. Exhale-focused breathing is a reasonable choice here, since five minutes of it measurably reduced physiological arousal in a randomised trial.
Then the same order tomorrow. The order matters more than the contents, because you are building a cue.
What about meditation before bed?
Fine if you enjoy it, but do not expect it to be the thing that fixes your sleep.
The largest systematic review of meditation programmes, covering 47 trials and 3,515 participants, found low evidence of no effect, or insufficient evidence of any effect, on sleep (Goyal et al. 2014, JAMA Internal Medicine). It found genuine moderate effects on anxiety and depression, which may help indirectly if those are what keep you up.
The point is one of ranking. Light, screens and caffeine have direct measured effects on sleep. Meditation's evidence sits elsewhere, which is worth knowing before you conclude you are bad at it. More on that in meditation myths.
Frequently asked questions
How long before bed should I stop looking at screens?
An hour is a reasonable target, though the evidence is about exposure rather than a precise cutoff. Reading on a light-emitting device before bed lengthened time to fall asleep, suppressed melatonin, delayed the circadian clock and reduced next-morning alertness compared with print (Chang et al. 2015). Putting the phone out of reach handles both the light and the content.
Does dimming the lights actually help you sleep?
Substantially. Compared with dim light, ordinary room light before bedtime produced a later melatonin onset in 99% of people and shortened melatonin duration by around 90 minutes (Gooley et al. 2011). Switching from overhead lighting to lamps for the last hour is one of the cheapest interventions available.
How late can I drink coffee?
Six hours before bed is the evidence-based cutoff. A 400 mg dose taken 6 hours before bedtime still significantly disrupted sleep relative to placebo, measured objectively (Drake et al. 2013). Individual sensitivity varies, so if you are affected, moving your cutoff earlier is reasonable.
Does a wind-down routine really work, or is it just habit?
Both, and the habit part is a feature. A routine lowers arousal directly and, repeated consistently, becomes a learned cue so the first step starts making you sleepy on its own. That is why a dull routine you actually repeat outperforms an elaborate one you do occasionally.
Will meditating before bed help me sleep?
Possibly indirectly, but the direct evidence is weak. The largest review found low or insufficient evidence of an effect of meditation programmes on sleep, while finding moderate evidence for anxiety and depression (Goyal et al. 2014). Light, screens and caffeine timing are the more direct levers.
Knit shows patterns, not diagnoses. Talk to a clinician about readings that concern you.
