Which habits change which metric

A table of common sleep habits against the numbers your wearable shows, with the direction of each effect, how sure the research is, and where nobody has looked.

• Evidence: Some
• Low effort
• Free
• Works: tonight to 2 weeks

Why it matters

If your app shows low HRV or little deep sleep, you want to know which habit to change. This table matches each habit to the metrics a study actually measured, with the direction of the change and how sure we are.

Every metric also has its own page that lists all the habits on this site that change it: time to fall asleep, deep sleep, REM, HRV, resting heart rate, efficiency and wake-ups, timing and total sleep. Those lists update automatically as pages are added.

How to do it

1. Find the metric you want to improve in the table’s top row.
2. Read down the column for habits marked “better” or “worse” with a confidence word.
3. Change one habit at a time for at least a week, so you can tell what worked.
4. Open the habit’s page for the how-to.

Words in brackets are our confidence: strong (meta-analysis or several trials), some (a few trials or large observational studies), weak (one small study, or mixed results). “Not studied” means we found no study measuring it.

Common mistakes

• Reading “not studied” as “no effect”. It means nobody has measured it, not that nothing happens.
• Expecting the full effect from one night. Light and wake-time habits work through your body clock and take days to weeks.
• Changing five things at once. You won’t know which one moved the number.

Effect on your sleep score

The biggest measured effects in healthy adults:

• Caffeine: 45 minutes less total sleep, 7% lower efficiency, 9 minutes longer to fall asleep and about 11 minutes less deep sleep on average [4].
• Alcohol: even about two standard drinks cut REM sleep, and the effect grows with dose [5]. In 4,098 employees, overnight heart rate was 1.4 to 8.7 beats per minute higher after drinking, depending on the amount [6].
• Screens: in 12 adults reading on a light-emitting e-reader for five evenings, falling asleep took about 10 minutes longer, REM was about 12 minutes shorter, and the body clock ran about 1.5 hours later than after a printed book [9].
• Warm bath: water at 40–42.5 °C (104–108.5 °F), 1–2 hours before bed, for as little as 10 minutes, shortened time to fall asleep [16].
• Exercise: in 14,689 WHOOP users, sessions ending at least 4 hours before sleep showed no link with worse sleep, whatever the intensity [15].

Evidence

Rated Some, because it varies by row. Caffeine and alcohol rest on meta-analyses [4][5]. Late meals are genuinely mixed: one trial of 20 adults found no change in sleep stages [10], while another of 13 young women found shorter, more broken sleep [11]. The nap study measured deep sleep intensity in young men in a lab [7], and the nap result for efficiency is from older adults wearing wrist monitors [8]. The temperature row comes from a review of lab studies [13]. The light rows rest on one lab study of the body clock [2] and one field study of 20 people [3], and the wake-time row on a study of 61 students [1].

Limits: very few of these studies used consumer wearables, and two authors of the large exercise study work for WHOOP [15].

General information for healthy adults, not medical advice. If you have a sleep problem that lasts, read when to see a doctor. Sources checked 4 October 2026. Spot a mistake? Tell us.