|

,

|

is your watch coaching you or just keeping score?


is your watch coaching you or just keeping score?

Russell D & Associates

The Performance Brief

Edition No. 014 · October 2026

Is your watch coaching you, or just keeping score?

Six things your wrist would like you to believe, checked against the trials. One holds, one half holds, and four would struggle in court.

9 minute read · For leaders who wear the data, and the operators who sell it

“Your sleep score is 58,” says the watch at 6.12am.

You felt fine until about 6.11.

By nine you are in a board meeting, a little slower than usual and quite sure it is because of the 58. There is research on exactly that moment, and we will get to it. First, some context on how many of us are wearing the thing that said it.

The American College of Sports Medicine says nearly half of US adults now own a fitness tracker or smartwatch. Its annual survey has named wearable technology the top fitness trend for three years running. The same report adds, with admirable restraint, that “rapid innovation often outpaces validation.”

So here is some validation. Six things the wrist promises, a plain verdict on each, and what the trials found when somebody actually checked.

The claimVerdict
It gets you movingHolds, then fades
More steps on the watch means better healthHalf true
10,000 steps is the targetWrong
The calorie number is what you burnedWrong
A bad sleep score just reports a bad nightWrong
A quiet watch means a healthy heartWrong

1 · “It gets you moving”

Verdict · Holds, then fades

This is the part the industry gets right. A review of 39 earlier reviews, covering about 164,000 people, found wearers walked about 1,800 more steps a day than non-wearers, roughly 40 extra minutes of walking, and lost about a kilogram (Ferguson and colleagues, Lancet Digital Health, 2022). The newest and largest umbrella review, published this year, agrees that activity goes up (Glisic and colleagues, Physiological Reviews, 2026).

Then the novelty wears off. In Ferguson’s data the extra had dropped to about 1,100 steps a day by four to six months, and to under 500 in studies that ran as long as four years. Your watch still congratulates you for walking to the fridge. It just has less to congratulate.

There is a quieter cost as well. Jordan Etkin ran six experiments with about 900 people for the Journal of Consumer Research in 2016. Tracking made people walk further and read more, and enjoy both less. When the tracking stopped, they did less than people who had never been tracked at all. Her conclusion deserves a place on the box. “For activities people do for fun, it may be better not to know.”

2 · “More steps on the watch means better health”

Verdict · Half true

Steps go up. The things your doctor measures mostly do not. Glisic’s review drew on 95 randomised trials and found the effects on heart and metabolic markers, quality of life, depression, anxiety and pain “limited” and “inconsistent”.

Singapore ran the cleanest test. The TRIPPA trial gave 800 office workers from 13 companies a Fitbit for a year (Finkelstein and colleagues, Lancet Diabetes & Endocrinology, 2016). The Fitbit on its own made no reliable difference to activity at six months. Cash rewards did, until the cash stopped and the gains left with it. Nobody in any group improved on weight or blood pressure at any point.

Then there is the result nobody in the category likes to mention. The IDEA trial put 470 adults through a weight-loss programme and gave half of them a wearable after six months (Jakicic and colleagues, JAMA, 2016). Two years on, the wearable group had lost 3.5 kg. The group without one had lost 5.9 kg. The gadget worked out, in effect, as a 2.4 kg handicap.

So why half true and not wrong? Because of PACE-UP. In the UK, 1,297 adults aged 45 to 75 were given a 12-week walking programme through their family doctor, with a basic pedometer and support from a practice nurse (Harris and colleagues, PLOS Medicine, 2019). Four years later they had fewer heart attacks and strokes and fewer fractures than people who did not take part. Roughly one heart attack or stroke was avoided for every 60 people put through it, and one fracture for every 28. Event numbers were small, and diabetes, depression and falls did not change.

Look at what that programme was. A cheap step counter, a short structured plan and a person checking in. Ferguson found the same thing across the field, with a device plus a programme working about half as well again as the device alone. The watch measures. A human gets the change.

3 · “10,000 steps is the target”

Verdict · Wrong

The number comes from a box, not a lab. In 1965, riding the enthusiasm left by the Tokyo Olympics, the Japanese company Yamasa sold a pedometer called the manpo-kei, which translates as “10,000-step meter”. It was a sales number that became a health target, and it has made a lot of people feel like failures at 7,400.

The best current evidence is kinder. Melody Ding and colleagues pooled 57 studies for Lancet Public Health in 2025. Against 2,000 steps a day, people walking 7,000 had about half the risk of dying during the follow-up. Their risk of dementia was about four in ten lower and of heart disease about a quarter lower. Most of that benefit had arrived by 5,000 to 7,000 steps. Beyond 7,000, the extra gains for most outcomes were modest.

Two honest limits. These studies followed people rather than assigning them steps, so they show a pattern, not proof of cause. And notice what they measured. Nobody has run a trial in which owning a wearable made people live longer. The longevity data is about walking, and walking does not care what counted it.

4 · “The calorie number is what you burned”

Verdict · Wrong

Of everything on your wrist, this is the least trustworthy figure. Stanford tested seven watches on 60 people in 2017 (Shcherbina and colleagues). Heart rate was mostly fine. Not one device estimated calories burned to within 20 in 100.

Nine years of product launches later, a 2026 review of 47 studies found recent devices still off by anywhere from a tenth to two fifths on calories, while step counts are now usually within 5 in 100 (Hogue and colleagues, Frontiers in Digital Health). Another 2026 study of four watches on 58 adults found the calorie error grew as body fat rose, on every brand tested (Kostrna and colleagues, PLOS One).

Read those two findings together. The people most likely to use the calorie number to plan a diet get the least accurate one. If you have ever “earned” a pastry from your watch, you were paid in a currency nobody audits.

5 · “A bad sleep score just reports a bad night”

Verdict · Wrong

Back to 6.12am. Start with what the device can see. A 2024 comparison at Brigham and Women’s Hospital, funded by Oura, put three trackers against a full lab sleep study for one night with 35 people. Agreement on sleep stages was moderate at best. The Apple Watch overestimated deep sleep by about 43 minutes on average, and failed to record six of the 35 people at all. Wearables are decent at telling you how long you slept and poor at telling you how well.

Now the part that surprises people. Being told you slept badly makes the next day worse, whatever actually happened. In one study, 164 students were shown a fake brain-scan readout saying their REM sleep had been above or below average (Draganich and Erdal, 2014). The “below average” group scored lower on attention and verbal tests. How well they themselves felt they had slept made no difference. Only what they were told mattered.

A smaller study rigged a bedroom clock (Rahman and colleagues, 2020). People who slept eight hours but were told five reacted more slowly the next day. People who slept five but were told eight reacted faster. It was 16 people, so treat it as a strong hint. At Oxford, 63 people with insomnia wore a watch that gave them fake sleep feedback (Gavriloff and colleagues, 2018). The “poor sleep” group reported more fatigue and duller thinking by evening, even though objective attention tests did not differ.

Your watch is decent at telling you how long you slept and poor at telling you how well. Being told you slept badly is enough to blunt the next day.

Does the score ruin the next night as well? The evidence is thin and it depends who you are. Healthy sleepers handed a one-off bad score showed no change in worry or sleep over the following week (Robson and colleagues, 2022). The trouble concentrates in people who already sleep badly or worry about it. Northwestern researchers coined the word “orthosomnia” in 2017 for patients who arrived self-diagnosed by their tracker and trusted it over the clinic’s own sleep study (Baron and colleagues). Somewhere between 3 and 14 adults in 100 fit that profile, depending on how strictly it is defined (Jahrami and colleagues, 2024).

The same data can help when someone reads it with you. In the NITE trial, 113 adults with insomnia got their wearable data plus guidance on how to interpret it, and their insomnia improved more than with standard sleep advice (Spina and colleagues, 2023). The most damaging thing your watch does to your sleep may be the message at 6.12am, not anything it noticed in the night.

6 · “A quiet watch means a healthy heart”

Verdict · Wrong

Apple’s hypertension notifications were cleared by the US Food and Drug Administration in September 2025. Compared with home blood pressure monitoring, the feature flagged 41 in every 100 people with high blood pressure. It missed the other 59. The American Heart Association’s journal Hypertension warned that clearance “should not be mistaken for accuracy or clinical readiness.”

A loud watch can earn its keep. In the Apple Heart Study, 419,297 people wore the watch and about 1 in 200 got an irregular-pulse alert (Perez and colleagues, NEJM, 2019). Among those who then wore a confirmatory ECG patch, about a third had atrial fibrillation confirmed. That is a real catch for a device on a wrist.

Alerts carry their own bill. Among 172 heart patients with atrial fibrillation, those who used wearables were more preoccupied with symptoms, worried more about treatment and used more tests and procedures (Rosman and colleagues, 2024). One in five felt intense anxiety when an alert fired. The study followed patients rather than assigning watches, so it shows a link, not a cause.

The line between wellness gadget and medical device is thinner than the marketing suggests. In July 2025 the FDA sent WHOOP a warning letter over a blood pressure feature promoted as “medical-grade”. The case closed in June 2026 after WHOOP changed the feature and its labelling. A notification is a reason to see a doctor. Silence is not a reason to skip one.

If your club sells the wrist

Every club owner has been pitched a heart-rate belt with a retention number attached. Myzone’s homepage promises a 25 in 100 increase in member retention and 35 in 100 more visits. Its own FAQ, further down the same page, says “up to” 25 and “up to” 33. The visit figure shrank by two points in the time it took to scroll, and neither carries a source or a method. That is normal for vendor claims, which is the point. They are sales figures, and they belong next to the manpo-kei.

The trials point somewhere more useful. The gadget on its own barely moves health. The gadget plus a person who reads it with you does better, in the walking data and in the sleep data alike. A belt is a commodity that arrives tomorrow from any online store. A coach who sits down with a member each month and turns the numbers into one decision is the part the evidence supports, and it is the part a member cannot buy elsewhere.

A test for every number on your wrist

One more finding before the test. The raw heart rate variability signal on the better rings comes close to a medical ECG (Dial and colleagues, 2025). The “readiness” score built on top of it is the manufacturer’s own recipe, with very little independent testing. The signal can be good while the verdict printed on top of it is a guess.

So tonight, go through the screens on your watch and ask one question of each number. If this were 20 points worse tomorrow, what would I do differently?

If the answer is a real decision, such as walking at lunch or booking a check-up, keep it on the home screen. If the honest answer is “feel worse and carry on”, hide it. Most people end up keeping steps and heart rate and quietly retiring the scores. That is roughly where the trials would put them too.

Over to you. You probably check your step count more often than you check the numbers that run your business.

Take the Performance Diagnostic A four-layer readout for the business, built to be acted on.

References

The figures above are these sources in plain terms. Sample sizes and study counts are as the researchers gave them.

  1. Ferguson T et al. Effectiveness of wearable activity trackers to increase physical activity and improve health. Lancet Digital Health, 2022. 39 systematic reviews, about 164,000 participants. About 1,800 steps a day more, about 40 minutes more walking, about 6 minutes more moderate to vigorous activity, about 1 kg weight loss. Effect about 1,100 steps a day at 4 to 6 months, under 500 at up to 4 years. 34 of 39 reviews rated critically low quality.
  2. Glisic M et al. Umbrella review of wearable activity trackers. Physiological Reviews, 2026. 39 systematic reviews, 98 original studies, 95 of them randomised trials. Effects on cardiometabolic markers, quality of life, depression, anxiety and pain described as limited and inconsistent. Over 80% of trials bundled the device with a coaching programme.
  3. Etkin J. The hidden cost of personal quantification. Journal of Consumer Research, 2016, 42(6), 967 to 984. Six experiments, about 900 participants.
  4. Finkelstein EA et al (TRIPPA). Lancet Diabetes & Endocrinology, 2016. 800 employees, 13 Singapore organisations, 12 months. Fitbit-only arm not significantly different from control on activity at 6 months. No improvement in health outcomes in any arm.
  5. Jakicic JM et al (IDEA). JAMA, 2016, 316, 1161 to 1171. 470 adults. Weight loss at 24 months 3.5 kg with wearable against 5.9 kg with standard intervention.
  6. Harris T et al (PACE-UP and PACE-Lift). PLOS Medicine, 2019. 1,297 adults aged 45 to 75. Fewer cardiovascular events and fractures at 4 years. Number needed to treat about 60 for cardiovascular events and about 28 for fractures.
  7. Ding D et al. Daily steps and health outcomes. Lancet Public Health, 2025. 57 studies, 35 cohorts. 7,000 against 2,000 steps a day associated with 47% lower all-cause mortality, 38% lower dementia, 25% lower cardiovascular disease, 22% lower depressive symptoms. Observational.
  8. Yamasa manpo-kei, launched 1965 in Japan, origin of the 10,000 step figure.
  9. Shcherbina A et al. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure. Journal of Personalized Medicine, 2017. Seven devices, 60 participants. No device estimated energy expenditure within 20% error.
  10. Hogue T et al. Systematic review of wearable accuracy. Frontiers in Digital Health, 2026. 47 studies. Energy expenditure error 10% to 40%, step counts usually within 5%.
  11. Kostrna J et al. PLOS One, 2026. 58 adults, four devices. Energy expenditure error increased with body fat percentage on all devices.
  12. Brigham and Women’s Hospital sleep tracker comparison. Sensors, 2024. Funded by Oura. 35 participants, one night against polysomnography. Apple Watch overestimated deep sleep by about 43 minutes and failed to record 6 of 35.
  13. Draganich C, Erdal K. Placebo sleep affects cognitive functioning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 2014. 164 participants.
  14. Rahman SA et al. Journal of Psychosomatic Research, 2020. 16 adults, manipulated clock time. Exploratory.
  15. Gavriloff D et al. Sham sleep feedback delivered via actigraphy. Journal of Sleep Research, 2018. 63 participants with insomnia. Objective attention measures did not differ between groups.
  16. Robson AR et al. Sleep and Biological Rhythms, 2022. 54 healthy sleepers, one-off sham sleep score.
  17. Baron KG et al. Orthosomnia. Journal of Clinical Sleep Medicine, 2017.
  18. Jahrami H et al. Brain Sciences, 2024. 523 adults. 3% to 14% met an orthosomnia profile depending on cut-off.
  19. Spina MA et al (NITE). Sleep, 2023. 113 adults with insomnia.
  20. Perez MV et al (Apple Heart Study). New England Journal of Medicine, 2019. 419,297 participants. About 0.5% notified. Atrial fibrillation confirmed in about 34% of those who returned an ECG patch.
  21. Apple Watch for Hypertension Screening. Hypertension (American Heart Association), 2025. doi:10.1161/HYPERTENSIONAHA.125.26031. Sensitivity about 41% against home blood pressure monitoring.
  22. Rosman L et al. Journal of the American Heart Association, 2024. 172 patients with atrial fibrillation. Observational.
  23. US FDA warning letter to WHOOP Inc, 14 July 2025, and close-out letter, 17 June 2026.
  24. Dial MB et al. Physiological Reports, 2025. 13 adults, 536 nights. Heart rate variability against ECG.
  25. ACSM Worldwide Fitness Trends 2026, American College of Sports Medicine.
  26. Myzone homepage and FAQ, myzone.org, accessed 21 October 2026. Homepage 25% retention and 35% visit frequency. FAQ up to 25% retention and up to 33% visit frequency. No source or method given.

Russell D & Associates

Executive Coaching & Consultancy

The Performance Brief is a weekly letter from Russell D & Associates for people who run clubs, teams and businesses.

www.russelldassociates.com


Leave a Reply

Your email address will not be published. Required fields are marked *