Russell D & Associates
The Performance Brief
Edition No. 012 · October 2026
The Cup You Never Feel
The coffee that costs you most is the one you are sure does nothing. A controlled trial took more than an hour of sleep from people who never noticed it was gone.
Three o’clock, every day
In Jakarta it is the kopi at three, when the afternoon slump meets the second round of meetings. In Jeddah, the meetings that mattered often started after evening prayers, and the Arabic coffee kept coming until you gave the cup a small shake to say enough. In every head office I have worked in, somebody makes a fourth coffee at four o’clock before the board pack.
Ask any of them how they sleep and most will tell you the same thing. “Coffee doesn’t affect me. I can drink it at ten at night and sleep like a baby.”
They believe it. In one controlled trial, the people who believed it were losing more than an hour of sleep a night.
The coffee that costs you most is the one you never feel.
1. The trial where nobody noticed
In 2013, Christopher Drake and colleagues at the Henry Ford Hospital sleep centre in Detroit ran a simple test on 12 healthy, normal sleepers. For four days, each person swallowed three identical pills a day: six hours, three hours and zero hours before their usual bedtime. On three of the days, one of the pills held 400 milligrams of caffeine, roughly two to three cups of coffee. On the fourth day, all three were dummies. Nobody knew which day was which.
A monitor measured their sleep at home. A diary recorded how they thought they had slept.
At bedtime and three hours before, the caffeine wrecked sleep, which surprised nobody. The result that matters came from the pill taken six hours out, the equivalent of a coffee at 4.30pm for a 10.30pm bedtime. Measured sleep fell by more than an hour. The diaries barely registered it. Most of the participants did not notice the night had been cut short.
Drake’s own summary was that people are less likely to detect caffeine’s effect on their sleep when they take it in the afternoon.
That is the whole problem in one line. You cannot manage a cost you cannot see.
2. What the bigger picture says
Twelve people is a small study, and 400 milligrams is a big dose. So here is the larger view.
In 2023, Carissa Gardiner and colleagues pulled together 24 caffeine and sleep studies for Sleep Medicine Reviews, and combined the sleep-time results from 20 of them, covering 340 people. Compared with the same people on a night without caffeine:
| What changed after caffeine | On average |
|---|---|
| Total sleep | 45 minutes less |
| Time to fall asleep | 9 minutes longer |
| Time awake during the night | 12 minutes more |
| Deep sleep | 11 minutes less |
Each row compares people with themselves on a caffeine-free night.
Run that across a working week. If the average held every weeknight, five nights at 45 minutes is close to four hours of sleep a week, paid for in coffee and never noticed. Most of it comes out of the night after a long day, when you needed it most.
The same team worked out how early you would need to stop for the effect to wash out. For a standard cup of coffee, about 107 milligrams, the answer was at least 8.8 hours before bed. For a pre-workout drink, roughly double the caffeine, it was more than 13 hours. For black tea, they found no cut-off at all.
The reason is simple chemistry. Caffeine takes around five hours, on average, for your body to clear half of it, and in some people closer to ten. The 4pm coffee is still half there when you switch off the light.
Nobody wakes up tired and blames the kopi they had at three.
3. The morning rule nobody tested
You may have been told the real mistake is drinking coffee too early. Wait 90 minutes after waking, the advice goes, or you will crash in the afternoon. It spread across podcasts and feeds in 2022 and it is now repeated as fact in most offices I walk into.
In 2024, a review by 20 sports nutrition researchers, led by Jose Antonio at Nova Southeastern University, looked for the evidence. They found none. Nothing shows that caffeine on waking causes an afternoon crash, and the 90 minute delay has never been directly studied. Michael Grandner, who runs the sleep and health research programme at the University of Arizona, put it plainly in 2025: there are no studies on the best time of day to drink it.
Here is the irony. The rule pushes your caffeine later into the day. The measured cost sits at the other end, in the afternoon cup. People have been worrying about the cup that carries no proven cost, and ignoring the one that does.
4. Where the evidence stops
I want you to use this, so here is what it cannot tell you.
The trial was small. Twelve people, one dose, four nights. It was well designed: dummy pills, people compared with themselves, sleep measured by a device rather than guessed. It is still twelve people, and 400 milligrams is more than most single coffees.
The review was mostly lab studies on light coffee drinkers. The people in them were healthy adults who drank low to moderate amounts. If you drink six coffees a day, your body may have adapted, and nobody has tested enough multi-day intake to say how much. The authors also say the results may not hold for people over 65.
Bodies differ. That five hour average hides a wide range. Some people clear caffeine fast, some slowly, and genes the studies did not measure play a part. The nine hours below is a sensible line, not a law.
What holds across all of it is the direction. Afternoon caffeine takes sleep, and the people losing it are the least likely to know.
5. The Nine Hour Rule
Your last coffee lands at least nine hours before you plan to sleep.
Draw the line. Take your usual bedtime and count back nine hours. Bed at 10.30pm means the last cup by 1.30pm. Set a daily phone alarm at that time and label it “Last cup”. The review’s figure for a single coffee was 8.8 hours. Nine is easier to remember and gives you a margin.
Swap, do not stop. After the line, switch to black tea, decaf or water. Tea showed no measurable cut-off in the review. Energy drinks and pre-workout carry roughly double the caffeine of a coffee and need more than 13 hours, so treat them as breakfast only.
Score it on Friday. Hold the line for five weeknights. On Friday morning, rate how you woke from one to ten, and compare it with how you woke on Monday. Your own before and after beats any average in this edition.
The bottom line
Afternoon caffeine is one of the few performance costs you pay every day without being able to feel it. The best evidence says it takes the best part of an hour from a night’s sleep, and the people paying are usually sure they are not.
The evidence has limits, and I have shown you where they sit. Inside them is a cheap, clear move that needs no willpower at the moment it matters, because the decision is made by an alarm at 1.30pm, not by you at 4pm.
Your one action before Friday. Tonight, work out your line: bedtime minus nine hours. Set the alarm, label it “Last cup”, and when it goes off tomorrow, make the next drink tea.
Over to you. What time is your last coffee, and what time do you go to bed?
Run the Executive Reset Free, 30 minutes. It checks whether the person at the top has the capacity to drive the plan, and sleep is the first place that capacity leaks.References
The figures above are the researchers’ own results put into plain terms: minutes of sleep and hours before bed. The trial compared people with themselves on caffeine and dummy-pill days. The review pools mostly laboratory studies of healthy adults. The underlying statistics are below.
- Drake, C., Roehrs, T., Shambroom, J. & Roth, T. (2013), Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed, Journal of Clinical Sleep Medicine, 9(11). 12 healthy normal sleepers, randomised, double-blind, placebo-controlled, within-subject; 400 mg caffeine at 0, 3 or 6 hours before habitual bedtime, placebo at the other times, plus an all-placebo day. Sleep measured by in-home monitor and diary. Caffeine at all three times significantly disrupted sleep; at 6 hours, objectively measured total sleep time fell by more than one hour, while subjective reports suggested participants were largely unaware of the disturbance.
- Gardiner, C., Weakley, J., Burke, L. M., Roach, G. D., Sargent, C., Maniar, N., Townshend, A. & Halson, S. L. (2023), The effect of caffeine on subsequent sleep: a systematic review and meta-analysis, Sleep Medicine Reviews, 69, 101764. 24 studies; total sleep time pooled across 20 (37 effect sizes, n = 340). Total sleep time reduced by 45.3 min (95% CI 29.0 to 61.5; p < 0.001); sleep onset latency +9.1 min; wake after sleep onset +11.8 min; deep sleep (N3 and N4) down 11.4 min; light sleep (N1) +6.1 min; sleep efficiency down 7.0%; no significant effect on REM. Cut-offs: coffee (107 mg per 250 mL) at least 8.8 h before bed; pre-workout (217.5 mg) at least 13.2 h; black tea (47 mg per 250 mL) none identified. Healthy adults, low to moderate habitual intake.
- Antonio, J., Newmire, D. E., Stout, J. R. and colleagues (2024), Common questions and misconceptions about caffeine supplementation: what does the scientific evidence really show?, Journal of the International Society of Sports Nutrition, 21(1), 2323919. No evidence that caffeine on waking causes an afternoon crash. Mean half-life about 5 hours, range roughly 1.5 to 9.5 hours.
- Grandner, M., quoted by the University of Arizona Department of Psychiatry (2025), on the absence of studies on optimal caffeine timing. Sleep Foundation (reviewed July 2025): half-life 2 to 12 hours, 4 to 6 hours in most cases.

Leave a Reply