You closed the laptop at 7:15. By any reasonable definition the workday ended there.

Then you loaded the dishwasher while rehearsing a conversation with a supplier. You half-watched something at nine and priced a renewal in your head through most of it. At 11:30 you got into bed, and the business had been running in the background for four hours without you touching a keyboard once.

Nobody would call that working. It is also nothing like being off.

The sleep guide names the routine that closes the gap. Capture, Close, Downshift. This page runs the whole thing: when to start it, what actually goes in the capture, how to keep it short enough that you will still run it on a bad night, what to do when you cannot stop at a fixed time, and how to tell after two weeks whether it did anything.

Read the boundary first. This is a way to end a workday. It is not a treatment for insomnia, and the difference matters enough that I come back to it at the end.

What you are actually trying to change

The thing with evidence behind it is called psychological detachment: being mentally away from work during off-hours, rather than merely being away from the desk.

A meta-analysis covering 86 publications, 91 samples, and 38,124 employees found detachment associated with better sleep (r = 0.30), lower fatigue (r = −0.42), and lower burnout exhaustion (r = −0.36).[1]

Read those numbers carefully. Most of the included studies were cross-sectional, heterogeneity ran moderate to high, and most samples came from Europe. Correlations of that kind cannot tell you which way the arrow points for any individual. Poor detachment may worsen sleep, poor sleep may weaken detachment, and a punishing quarter may drive both.

What the finding does establish is that the thing you are chasing has a name and a measurable relationship with how you sleep and how tired you feel. Your evening is a variable rather than a mood.

There is also modest evidence that detachment can be learned. A recovery training study ran two sessions a week apart with 48 people against 47 on a waitlist. Recovery experiences and self-efficacy rose, self-reported sleep quality improved at both follow-ups, and perceived stress fell by week three.[2]

Ninety-five people, self-reported outcomes, no entrepreneurs, and the training produced no effect at all on emotional exhaustion. Treat it as encouragement that the skill responds to practice, well short of proof that a routine fixes an evening.

Run it 60 to 90 minutes before you want to be asleep

Most shutdown advice never says when, which is exactly why most people never run it.

The routine belongs at the end of the workday rather than the end of the evening, and it belongs at your desk. Running it in bed defeats the purpose, because the point is to put distance between the last work thought and the pillow.

Sixty to ninety minutes is a practical choice rather than a tested dose. No trial has established a correct gap. What the timing buys you is a stretch of evening that sits after work has been formally closed, which is the part most entrepreneurs have never actually had.

The whole routine should take eight to twelve minutes. Anything longer becomes a fifth task, and the first genuinely hard day will kill it.

Capture what is still open

Write down the work your mind is trying to hold, in enough detail that you trust yourself to restart tomorrow.

Specificity is doing the work here. A small polysomnographic study asked 57 healthy adults aged 18 to 30 to write for five minutes before bed. Those assigned to write a specific to-do list fell asleep faster than those who wrote about completed activities, and within that group more specific lists were associated with faster sleep onset.[3]

One night, one sleep lab, young healthy adults, and nobody running a company. Take the direction and hold the size of the effect loosely.

The format that works looks like an instruction to a competent stranger:

Send Maya the revised proposal by 10:00. Confirm the page-four price with Dana first. Her number is in the thread from Tuesday.

Compare that with a line reading only "proposal, Maya." The second version leaves three questions unanswered, and your mind will keep raising all three at midnight until it gets answers.

Here is where the popular advice overreaches. A four-arm randomised controlled trial followed 186 participants over 14 days testing worry postponement, with and without added if-then planning. The augmented version cut roughly 15 minutes off daily worry duration compared with standard postponement. Neither version produced a significant improvement in any sleep outcome against control.[4]

So the writing has decent support for reducing worry and thin support for improving sleep. Fifteen fewer minutes of worrying is worth eight minutes at a desk on its own terms. Expecting those same eight minutes to deliver sleep sets you up to abandon a useful habit for failing at something it was never shown to do.

Close the business for the night

Capture empties your head. Closing decides that the day is over, which for someone who owns the business is a genuinely different act.

An employee's workday ends when they leave. Yours ends when you say it does, and if you never say it, it does not.

Pick the first action for tomorrow, name when you will return to it, and shut the inbox and the tabs that invite one more round. Then say the closing line out loud or write it down:

Work is closed. First task tomorrow is the proposal review at 9:30.

That sentence has never been tested as a sleep method and I would not present it as one. Its job is narrower. It gives your mind a decision it can point at, so the question "should I be doing something about work right now" stops arriving at 8:40 and 9:15 and 10:05 with no answer available.

If a task keeps pulling at you, leave a restart line the way you would before a deep-work block. If what is pulling is an unresolved decision rather than an unfinished task, capture the question and put it in a proper thinking window tomorrow, which the mental energy guide covers as a placement problem.

Downshift into lower input

Move into something with less light, less incoming information, and fewer decisions to make.

Dim the room. Set out what you need for the morning. Read something undemanding, take a warm shower, put the phone somewhere that requires standing up.

Keep the order fixed and short. You will only ever run this when you are tired, so it has to survive being run by a tired person with no motivation and no interest in doing it well.

Judge the downshift on whether input went down and work stopped. Do not judge it on whether you fell asleep quickly, because that turns a wind-down into a performance with a pass mark, and effort is the one thing that reliably makes sleep harder.

The shutdown that quietly becomes another task

Here is the most common way this fails, and it rarely looks like failure.

The routine grows. Capture becomes a full review of tomorrow. The closing line becomes a planning session. Someone adds journaling, then a gratitude list, then a stretch sequence, and within three weeks the wind-down runs 40 minutes and requires more decisions than the workday it was meant to end.

A systematic review of pre-sleep cognitive activity found people with insomnia reported more planning, problem-solving, worry, and mental arousal before sleep than good sleepers, with interventions aimed at those thought patterns producing results that ranged from helpful through negligible to harmful.[5] Some ways of working on your own thinking at bedtime make things worse.

The related pattern for worry and rumination points the same direction. A systematic review and meta-analysis found more worry and rumination associated with poorer self-reported sleep quality, less total sleep, and longer time before sleep, resting mostly on cross-sectional self-report.[6]

The practical guard is a hard cap on length. If the routine cannot be run in about ten minutes by someone who is exhausted and slightly annoyed, it is too big, and the part to cut is whatever you added most recently.

When you cannot stop at a fixed time

Every version of this advice assumes an evening you control. Founders with clients in other timezones, small children, or a business in its first two years control considerably less than that.

Anchor the routine to the end of work rather than to a clock. If work ends at 6:30 on Tuesday and 10:15 on Thursday, the routine still runs at the end of work. A fixed hour will break within a week and take the habit with it.

Give a genuine second shift its own shutdown. Plenty of entrepreneurs work again after the kids are down. That block is a workday, so it gets Capture and Close too, even if the whole thing takes four minutes.

Shorten before skipping. On the worst nights, capture alone is worth running. One specific line beats a perfect routine you did not do.

Decide in advance which nights are exceptions. A launch week, a filing deadline, or a family crisis will override this. Naming those in advance keeps an exception from turning into evidence that the routine does not work for you.

Watch the caffeine that props up a late shift. A systematic review found caffeine generally increased time to fall asleep and reduced total sleep, sleep efficiency, and reported sleep quality, with dose, timing, age, and individual sensitivity all changing the result.[10] No universal cutoff exists, which makes this worth testing on yourself over a week.

When tonight goes wrong anyway

Some nights the routine runs properly and you still end up awake at 3 a.m. with a receivable. The sleep guide holds the rule for that moment, and the short version is to avoid reopening the workday.

Getting out of bed when sleep will not come is the one piece of that with formal clinical backing. The American Academy of Sleep Medicine's guideline conditionally recommends stimulus control as a single-component therapy for chronic insomnia in adults, while giving its strong recommendation to multicomponent cognitive behavioral therapy for insomnia and advising against sleep hygiene alone as a treatment.[8]

Note what that means for the routine on this page. A shutdown sequence sits closer to sleep hygiene than to therapy, and the 2021 guideline explicitly declines to endorse hygiene alone as a treatment for chronic insomnia. For ending a workday it is a reasonable tool. For insomnia running three nights a week across three months it is the wrong instrument, and eight minutes at a desk does not substitute for the multicomponent therapy the guideline actually recommends.

How to tell whether it is working

Give it two weeks, and measure the right thing.

Each morning, write down three items: whether you ran the routine, roughly how long after getting into bed you think you fell asleep, and how many times work-related thinking pulled you back during the evening. That third one is the number most people never track, and it is the closest thing you have to a detachment measure.

Two weeks is enough to see a direction. A single night tells you nothing at all, since sleep varies for reasons that have no relationship to your evening.

Judge it on the evening rather than the night. If work-related thinking dropped from six intrusions to two while your sleep onset barely moved, the routine is doing precisely what the evidence supports and you should keep it. Grading it purely on how fast you fall asleep means grading it on the outcome the trials failed to move.

If a one- to two-week record shows a pattern you do not like, bring it to a clinician. A sleep diary is the artifact they will ask for anyway.

Tonight

  1. Set the trigger. When you close the laptop tonight, run Capture and Close at your desk before you leave it. Eight minutes.
  2. Write one capture line specific enough for a stranger to act on. Name the person, the deliverable, and the first step.
  3. Start the morning record tomorrow. Routine run, rough sleep onset, and how many times work pulled you back during the evening.

The routine runs on paper and costs nothing. If a guided audio track helps mark the downshift, Awakened Mind's Sleep Support may serve as one optional part of it. Keep the volume low, limit listening time, and remove headphones before sleep unless your device maker's current guidance says otherwise. Open Awakened Mind in the App Store.

If the evening ended in conflict, recover from that before asking yourself to run a quiet routine. If low energy is shaping the whole day rather than just the evening, that belongs to mental state management more broadly.

The business will still be there at nine. Closing it for the night is a decision you get to make, and nobody else is going to make it for you.

References

  1. A meta-analysis on antecedents and outcomes of detachment from work. Frontiers in Psychology, 2017. Link Limit: 86 publications and 38,124 employees, but mostly cross-sectional, heterogeneity moderate to high, and most samples European. Correlational, so direction of effect is unresolved.
  2. Learning how to recover from job stress: effects of a recovery training program. Journal of Occupational Health Psychology, 2011. Link Limit: 48 trained against 47 waitlisted, self-reported outcomes, and no effect on emotional exhaustion.
  3. The effects of bedtime writing on difficulty falling asleep: a polysomnographic study comparing to-do lists and completed activity lists. Journal of Experimental Psychology: General, 2018. Link Limit: 57 healthy adults aged 18 to 30, one night in a sleep lab. No entrepreneurs and no repeated measurement.
  4. Effects of worry postponement on daily worry and sleep: a randomised controlled trial. Psychology & Health, 2025. Link Limit: 186 participants over 14 days. Reduced worry duration, and no significant improvement in any sleep outcome against control.
  5. Pre-sleep cognitive activity in adults: a systematic review. Sleep Medicine Reviews, 2019. Link Limit: interventions on pre-sleep thinking ranged from helpful to negligible to harmful. No single reliable method identified.
  6. The association between worry and rumination with sleep in non-clinical populations: a systematic review and meta-analysis. Health Psychology Review, 2021. Link Limit: mostly self-report and cross-sectional. Cannot establish direction for any individual.
  7. Insomnia Treatment. National Heart, Lung, and Blood Institute, 2022. Link Limit: general patient guidance. Habits described may support sleep and are not a stand-alone treatment for long-term insomnia.
  8. Behavioral and psychological treatments for chronic insomnia disorder in adults: an American Academy of Sleep Medicine clinical practice guideline. Journal of Clinical Sleep Medicine, 2021. Link Limit: stimulus control carries a conditional recommendation as single-component therapy. The strong recommendation is for multicomponent CBT-I, and the guideline advises against sleep hygiene alone as treatment.
  9. Insomnia Diagnosis. National Heart, Lung, and Blood Institute, 2022. Link Limit: describes diagnostic criteria. Diagnosis requires a clinician reviewing history, health, and medications.
  10. Coffee, caffeine, and sleep: a systematic review of epidemiological studies and randomized controlled trials. Sleep Medicine Reviews, 2016. Link Limit: timing, dose, age, and sensitivity changed results. No universal cutoff time.