7,000 steps,
not 10,000.
A 2025 meta-analysis in The Lancet Public Health pooled 57 studies and drew the curve: how much lower the risk of nine outcomes was, step by step, versus people who walk about 2,000 a day. Most of the gain arrives long before 10,000. Put in your average and see what your next thousand is worth.
In the studies, people averaging ~4,500 steps a day, compared with the 2,000-step group (95% CI 30%–46%).
Next band, 5k–6k: about 42%. On the curve, your next 1,000 steps ≈ +5 points.
Moderate-certainty evidence (GRADE), 14 studies.
You're on the list.
One more question — it decides what we build first.
Noted — nothing is charged today. Payments open with the beta, and you'll be first to hear. Thank you.
The first 3,000 extra steps are worth the most
That is the mortality difference the studies found at 4,000 steps versus 2,000. Going from 7,000 to 10,000 added one more point. The curve is steep where sedentary people live and almost flat where fitness apps set their goals.
Where the all-cause mortality curve levels off (HR 0.53 at 7,000; 0.52 at 10,000). The authors call it "a more realistic and achievable target for some" — and note 10,000 remains a fine goal for people already active.
Dementia at 7,000 versus 2,000 — and unlike mortality, that curve keeps sloping past 7,000, as do depressive symptoms (−22% at 7,000, −33% at 10,000). Plateaus are outcome-specific, so the app shows each curve separately rather than one blended score.
The year a Japanese pedometer was named manpo-kei, "ten-thousand-step meter". The number came from the product name, not from a study — and it has been demotivating people at 3,000 to 5,000 ever since.
What Step ROI does
It reads your steps from Apple Health and shows the band you are in — a 14-day average, because a single day says nothing and a bad day should not cost you anything.
Your goal is your next band, not 10,000: a 14-day median plus a nudge, rounded to something you can hit. Past 7,000 the goal turns into holding it five days a week, with an optional stretch for the outcomes that keep improving.
Every Monday you get a short note in plain English: what moved, what it corresponds to on the curves, one thing to try this week. A lock-screen widget keeps the band where you can see it without opening the app.
Steps stay on your phone. Nothing is shared with advertisers or stored in iCloud.
The science, in full
Ding and colleagues reviewed 57 studies from 35 cohorts and pooled 31 of them. Steps were measured with worn devices (77% accelerometers, mostly hip-worn), outcomes were followed for years, and the reference group was people averaging 2,000 steps a day. The table is the whole result; the calculator above only reads off it.
| Outcome | 4,000 | 7,000 | 10,000 | Shape | Certainty |
|---|---|---|---|---|---|
| All-cause mortality 14 studies | 0.64 0.57–0.73 | 0.53 0.46–0.60 | 0.52 0.45–0.59 | nonlinear bend ≈ 5,391 | moderate |
| Cardiovascular disease (incidence) 6 studies | 0.87 0.81–0.93 | 0.75 0.67–0.85 | 0.70 0.62–0.79 | nonlinear bend ≈ 7,802 | moderate |
| Dementia 2 studies | 0.81 0.75–0.87 | 0.62 0.53–0.73 | 0.55 0.46–0.66 | nonlinear bend ≈ 8,829 | moderate |
| Depressive symptoms 3 studies | 0.91 0.88–0.93 | 0.78 0.73–0.83 | 0.67 0.61–0.75 | linear | moderate |
| Type 2 diabetes 4 studies | 0.94 0.89–1.00 | 0.86 0.74–0.99 | 0.78 0.61–0.99 | linear | moderate |
| Cancer mortality 3 studies | 0.72 0.63–0.82 | 0.63 0.55–0.72 | 0.60 0.51–0.70 | linear | moderate |
| Cardiovascular disease (mortality) 3 studies | 0.61 0.44–0.85 | 0.53 0.37–0.77 | 0.56 0.40–0.79 | linear | low |
| Falls 4 studies | 0.86 0.81–0.91 | 0.72 0.65–0.81 | 0.70 0.63–0.79 | nonlinear bend ≈ 8,846 | very low |
| Cancer (incidence) · not statistically significant 2 studies | 0.97 0.95–1.00 | 0.94 0.87–1.01 | 0.90 0.81–1.01 | linear | low |
Hazard ratios vs the 2,000-steps/day group, with 95% confidence intervals. Certainty is GRADE as rated by the authors. Source: Ding D, Nguyen B, Nau T, et al. Daily steps and health outcomes in adults: a systematic review and dose-response meta-analysis. Lancet Public Health 2025; 10: e668–81. Full text. Curve data version 2025.07-ding-v1.
What this can and cannot tell you
These are associations between groups in cohort studies, not a trial. People who walk 2,000 steps a day may differ in ways the studies could not fully remove — health, frailty, work — and the authors say so. Just under half the studies followed adults over 65; the relative pattern likely holds more widely, but your absolute risk at 30 is small in the first place, so the same percentage means fewer events.
The percentages compare groups to the 2,000-step group; they are not your personal probability of anything, and Step ROI never turns them into one. Cancer incidence did not reach significance and falls have very low-certainty evidence, so neither appears in the headline tabs.
Step ROI shows where your average daily steps fall on curves from published population studies. These are associations between groups of people, not predictions about you, and not medical advice. Percentages are relative to people averaging about 2,000 steps a day. If you have a health condition or are changing your activity a lot, talk to a doctor.
You're on the list.
One more question — it decides what we build first.
Noted — nothing is charged today. Payments open with the beta, and you'll be first to hear. Thank you.