Ten Thousand Steps Was Never a Real Number

A Number With a Birthday

Pull out a phone. Somewhere in its health app, almost certainly, sits a daily step goal, and almost certainly that goal is 10,000. Nobody set it for you personally. Nobody asked your age, your resting heart rate, or whether you have knees left to spare. The number simply arrived, pre-loaded, treated by the software and by most of the people using it as though it were a clinical recommendation with the same standing as a vaccine schedule or a cholesterol target.

It is not that. We went looking for the trial that established 10,000 steps a day as a health threshold, the way you can go looking for the trials behind “150 minutes of moderate activity a week” or “five servings of fruit and vegetables.” There isn't one. The paper that sent us down this path says so directly, in a sentence blunt enough that we read it twice: a 2019 cohort study of older women, published in JAMA Internal Medicine, states plainly that “the origin of the goal of 10,000 steps per day is unclear” and that it “likely derives from the trade name of a pedometer sold in 1965” [2]. That is not a rumor repeated by wellness bloggers. That is a peer-reviewed epidemiology paper telling its readers, in its own introduction, that the number everyone assumes it is about to validate has no such pedigree.

What the club actually checked

We did not take that sentence on faith. We traced the citation trail it points to, confirmed the company and the year independently, and then went to find out what a large, current, multi-outcome meta-analysis says the number should actually be. That paper is the real subject of this field note. The marketing history is the hook; the dose-response curve is the substance.

Manpo-Kei: What the Watch Company Actually Sold

The product was called the manpo-kei. In Japanese, man is 10,000, po is step, and kei is meter or gauge — a “10,000-step meter,” named for exactly what it displayed and nothing more [2]. It was made and sold in 1965 by Yamasa Tokei Keiki, a clock and instrument manufacturer, into a Japan that had just hosted the 1964 Tokyo Olympics and was riding a genuine, well-documented boom in public enthusiasm for fitness and walking.

Ten thousand is a good number for a product name. It is round in decimal, easy to say, easy to print on a dial, and large enough to sound aspirational without sounding impossible. None of those are qualities a dose-response curve is obligated to share. As far as the historical record traced by later exercise-science literature shows, nobody in 1965 ran a study comparing health outcomes across different step counts and found that the curve did something special at 10,000. The number came from a naming meeting, not a laboratory.

The Slogan Outgrows Its Packaging

A catchy product name is not, by itself, enough to become a durable piece of global health folklore. What happened next is the more interesting part of the story, and it is a slow one: walking clubs organized around the round number, and the figure kept getting picked up by each new generation of researchers and institutions as a convenient, pre-existing benchmark rather than a number anyone was newly proposing on the evidence.

By 2004, when the exercise scientists Catrine Tudor-Locke and David Bassett published the paper that gave the field its standard step-count categories — sedentary, low active, somewhat active, active, highly active — 10,000 steps a day was already so entrenched as the public's benchmark for “active” that the category system was built to accommodate it, explicitly as a widely recognized target rather than as a figure their own data had newly derived [3]. A follow-up paper four years later, co-authored with the Japanese physical-activity researcher Yoshiro Hatano, revisited the same question and reached a similar diagnosis: a good round number, adopted worldwide, resting on much thinner evidence than its ubiquity implied [4]. By 2023, a group of physical-activity researchers writing in the British Journal of Sports Medicine gave the whole history a name in their title — “Journey of a Thousand Miles: From Manpo-Kei to the First Steps-Based Physical Activity Recommendations” — treating the pedometer's marketing copy as the acknowledged starting point of an entire scientific literature that only later caught up to it [5].

Fifty Years Later, Someone Finally Checks

The first serious attempt to measure where the real curve bends, rather than argue about where marketing put it, came from a study of 16,741 women with a mean age of 72, followed for a mean of 4.3 years, during which 504 of them died [2]. Sorted into quartiles by median daily steps — 2,718, 4,363, 5,905, and 8,442 — their adjusted mortality hazard ratios came out at 1.00 (reference), 0.59, 0.54, and 0.42. The paper's own description of the shape is direct: mortality risk “progressively decreased before leveling at approximately 7,500 steps/day” [2]. Not 10,000. Not a hard wall either — a curve that keeps softening as it climbs and eventually goes nearly flat.

A second large cohort complicated the picture rather than closing it. A 2022 analysis of 78,500 UK Biobank participants, followed for a median of seven years, reported that benefits for mortality, cancer, and cardiovascular disease kept accruing up to 10,000 steps a day, and that the pace of stepping (not just the daily total) carried an independent protective association [6]. Two well-conducted cohorts, two different apparent ceilings. That disagreement is exactly the kind of thing a single study cannot resolve on its own, and it is exactly what a pooled, multi-cohort, multi-outcome review exists to adjudicate.

What the 2025 Meta-Analysis Actually Did

Ding and colleagues searched PubMed and CINAHL for prospective studies published between January 2014 and February 2025 that measured daily steps with a wearable device, rather than by self-report, in adult populations [1]. That device requirement matters: self-reported step counts are notoriously unreliable, and restricting the review to accelerometer- and pedometer-measured data is a large part of why this synthesis is more trustworthy than the folklore it is replacing.

Eight outcome domains is the point. Earlier work, including both cohort studies above, could only speak to whichever outcomes its one dataset happened to track. A review broad enough to cover mortality, cardiovascular disease, cancer, type 2 diabetes, dementia, depressive symptoms, falls, and physical function in the same analysis can ask a question neither of them could: does the curve look the same for every outcome, or only for some of them? The answer, as the next two sections show, is genuinely surprising.

Seven Thousand, Compared to Two Thousand

Every headline number in this paper is a comparison between people who take about 7,000 steps a day and people who take about 2,000. Relative to that 2,000-step baseline, 7,000 steps a day is associated with a 47% lower risk of all-cause mortality (hazard ratio 0.53, 95% CI 0.46–0.60), a 25% lower risk of cardiovascular disease incidence (0.75, 0.67–0.85), a 38% lower risk of dementia (0.62, 0.53–0.73), and a 22% lower risk of depressive symptoms (0.78, 0.73–0.83) [1]. Cardiovascular mortality and cancer mortality were each about 47% and 37% lower, type 2 diabetes incidence 14% lower, and fall risk 28% lower, though the review reports those last four without a pooled confidence interval as precise as the first four [1].

Does the Curve Actually Plateau?

Here is the part the popular coverage of this study tends to flatten out, and it is the part we found genuinely interesting. The review does not report one dose-response shape. It reports two.

For all-cause mortality, cardiovascular disease incidence, dementia, and falls, the paper describes “an inverse non-linear dose-response association, with inflection points at around 5,000–7,000 steps per day” [1]. That is a real plateau in the statistical sense: the curve keeps falling as steps increase, but its slope softens sharply past that range, so each additional thousand steps buys noticeably less than the thousand before it did. It is a diminishing-returns curve, not a cliff — nothing in the data suggests walking 9,000 steps instead of 7,000 does you harm, or that the curve goes flat as a rule rather than merely soft.

For cardiovascular disease mortality, cancer incidence, cancer mortality, type 2 diabetes incidence, and depressive symptoms, the paper reports something different: “an inverse linear association” [1]. No inflection point is described for these outcomes within the range the included studies covered. The practical reading is blunt: for this half of the outcome list, the data simply do not show a step count at which more stops helping. Seven thousand is a reasonable target if mortality and dementia are what you care about most. It is not obviously a ceiling if you're weighing cardiovascular or cancer mortality specifically.

Seven thousand steps is a genuinely good answer to “where does the mortality curve stop rewarding you much for walking more.” It is not an answer to “where should I stop caring about steps,” because five of the nine outcomes in this dataset never told the researchers where that second point was.

Reading the Uncertainty Honestly

Not every number in a meta-analysis deserves the same trust, and this one says so about itself. The review rates its evidence as moderate certainty for most outcomes, but explicitly low certainty for cardiovascular mortality, cancer incidence, and physical function, and very low certainty for falls [1] — four of the nine figures we just charted. Low certainty in this context typically means fewer pooled cohorts, wider intervals, and more room for the estimate to move once more studies exist. The cancer incidence result is the clearest illustration sitting right in the data: a hazard ratio of 0.94 with a confidence interval running from 0.87 to 1.01 is, by the conventional threshold, not distinguishable from no effect at all [1]. We are reporting that null result rather than quietly dropping it, because a field note that only shows you the significant rows is doing the same selective flattening we are criticizing in the marketing history.

There is a second, more structural caveat that no amount of pooling fixes. Every study feeding into this review is observational: it compares people who happen to walk different amounts, not people randomly assigned to walk different amounts. People who already have undiagnosed illness tend to walk less before anyone notices they are sick, which can make low step counts look more dangerous than walking itself explains. The review's authors cannot fully rule that out, and neither can we. Pooling 57 studies makes the estimate more precise; it does not, by itself, make the causal story airtight.

The Club's Arithmetic: Seventy Percent of a Slogan

None of the arithmetic in this section is the paper's. It is ours, done on top of the paper's headline number and a widely used, independently sourced walking-distance approximation, and we are labeling it that way throughout.

The fraction is the easy part. 7,000 steps is 7,000 ÷ 10,000 = 0.70, or 70%, of the marketing number everyone grew up with. The sixty-year-old slogan was asking for roughly three-tenths more daily walking than the best current pooled evidence says is doing most of the measurable work for mortality, cardiovascular incidence, dementia, and falls.

Turning that into distance takes one more assumption, and we are borrowing it rather than deriving it ourselves: a commonly cited average-adult-stride approximation puts 10,000 steps at roughly 8 kilometres, or about 5 miles. Scaling that figure by our 0.70 fraction — simple proportion, nothing the Lancet Public Health paper itself calculates — puts 7,000 steps at roughly 5.6 kilometres, or about 3.5 miles. The gap between the two targets, 3,000 steps, comes out to roughly 2.4 kilometres, or about 1.5 miles, of daily walking that the marketing number was demanding on top of what the pooled evidence associates with most of the benefit.

Why the Wrong Number Won for Sixty Years

It is worth pausing on why a number with no clinical origin outcompeted decades of more careful, more hedged research communication. Part of the answer is pure design: 10,000 is round, easy to remember, and works identically as a marketing spec or a software default regardless of what any given dose-response curve says, which made it frictionless for pedometer companies, and later phone and wearable manufacturers, to keep reusing without ever having to justify it clinically. Part of the answer is institutional: once the World Health Organization's actual physical-activity guidance was published, it was framed in minutes of moderate-to-vigorous activity per week, not steps at all [7] — there was never an official step-count guideline for journalists or app developers to cite instead, so the pre-existing folk number filled the vacuum by default.

And part of the answer is that overturning folklore this entrenched needed more than one good study. The 2019 cohort of older women was compelling, but it was one demographic, one country's health system, one relatively short follow-up window [2]. It took a review broad enough to span 57 studies and eight outcome domains before public health communicators had something with the breadth to plausibly replace a sixty-year-old slogan rather than merely footnote it.

What the Paper Does Not Settle

The authors do not claim to have retired 10,000 as a target. Their own stated conclusion is that “10,000 steps per day can still be a viable target for those who are more active,” alongside the finding that 7,000 is “associated with clinically meaningful improvements in health outcomes and might be a more realistic and achievable target for some” [1]. That is a considerably more modest claim than “science debunks 10,000 steps,” and it is the one we think the evidence actually supports.

Several open questions remain outside this review's reach. It cannot separate the contribution of walking speed or cadence from raw step count, a variable the 2022 UK Biobank study found to matter independently [6]. It pools studies across different devices, populations, and follow-up lengths, which is a strength for generalizability and a weakness for precision. And because every included study is observational, none of this establishes that walking more causes the outcomes it is associated with, only that the two reliably travel together across dozens of independent cohorts.

The Actual Takeaway

If you are walking closer to 2,000 steps a day than 7,000, the marginal value of your next thousand steps is large under this dataset, regardless of which round number you are technically chasing. If you are already well past 7,000, the mortality, cardiovascular-incidence, dementia, and fall-risk curves in this review suggest you have likely captured most of what device-measured step count alone predicts for those four outcomes, and chasing 10,000 specifically buys comparatively little more on them. If cardiovascular or cancer mortality is what worries you most, this particular dataset does not hand you a number at which to stop; those associations kept strengthening in a straight line across the whole range the included studies measured.

The number that started this whole field of study was never a finding. It was a product name that got lucky enough to sound plausible for sixty years. The number that should probably replace it in your head is not really 7,000 either — it is the shape of nine different curves, four of which bend and five of which don't, none of which particularly cares what a 1965 clock company decided sounded good on a dial.

References

1. 1.

   Ding, D., Nguyen, B., Nau, T., Luo, M., del Pozo Cruz, B., Dempsey, P. C., Munn, Z., Jefferis, B. J., Sherrington, C., Calleja, E. A., Chong, K. H., Davis, R., Francois, M. E., Tiedemann, A., Biddle, S. J. H., Okely, A., Bauman, A., Ekelund, U., Clare, P. & Owen, K. (2025). Daily steps and health outcomes in adults: a systematic review and dose-response meta-analysis. The Lancet Public Health 10(8), e668–e681. doi:10.1016/S2468-2667(25)00164-1

2. 2.

   Lee, I.-M., Shiroma, E. J., Kamada, M., Bassett, D. R., Matthews, C. E. & Buring, J. E. (2019). Association of step volume and intensity with all-cause mortality in older women. JAMA Internal Medicine 179(8), 1105–1112. doi:10.1001/jamainternmed.2019.0899

3. 3.

   Tudor-Locke, C. & Bassett, D. R. (2004). How many steps/day are enough? Preliminary pedometer indices for public health. Sports Medicine 34(1), 1–8. doi:10.2165/00007256-200434010-00001

4. 4.

   Tudor-Locke, C., Hatano, Y., Pangrazi, R. P. & Kang, M. (2008). Revisiting “how many steps are enough?” Medicine & Science in Sports & Exercise 40(7 Suppl 1), S537–S543. doi:10.1249/MSS.0b013e31817c7133

5. 5.

   Stamatakis, E., Ahmadi, M. N., Murphy, M. H., Chico, T. J. A., Milton, K., del Pozo Cruz, B., Katzmarzyk, P. T., Lee, I.-M. & Gill, J. M. R. (2023). Journey of a thousand miles: from ‘Manpo-Kei’ to the first steps-based physical activity recommendations. British Journal of Sports Medicine 57(19), 1227–1228. doi:10.1136/bjsports-2023-106869

6. 6.

   del Pozo Cruz, B., Ahmadi, M. N., Lee, I.-M. & Stamatakis, E. (2022). Prospective associations of daily step counts and intensity with cancer and cardiovascular disease incidence and mortality and all-cause mortality. JAMA Internal Medicine 182(11), 1139–1148. doi:10.1001/jamainternmed.2022.4000

7. 7.

   Bull, F. C., Al-Ansari, S. S., Biddle, S., Borodulin, K., Buman, M. P., Cardon, G., et al. (2020). World Health Organization 2020 guidelines on physical activity and sedentary behaviour. British Journal of Sports Medicine 54(24), 1451–1462. doi:10.1136/bjsports-2020-102955