Daily steps: what the research actually says, and how to track them honestly

Several large cohort studies over the last decade, following tens of thousands of adults with wearable step counters for years, converge on the same shape. Compared with the least active group — under about 4,000 steps a day — the risk of all-cause mortality falls substantially with each additional couple of thousand steps, and the drop is steepest at the bottom: going from 3,000 to 5,000 buys more than going from 8,000 to 10,000. The curve keeps improving up to roughly 7,000-8,000 steps a day in adults under sixty and somewhat lower, around 6,000-7,000, in older adults, after which additional steps still help but by progressively less. Above 10,000 the benefit in these studies is small and inconsistent. Step intensity — walking faster — adds a separate benefit on top of the count. So the honest target is not a single number: it is "more than you do now, up to around 7,500, and some of it brisk". The 10,000 target is not harmful; it is just arbitrary, and for someone at 3,000 it is discouraging in a way that 5,000 is not.

A phone step counter reads the accelerometer — the motion sensor that detects the phone's movement in three axes — and looks for the rhythmic, roughly one-to-two-per-second oscillation that walking produces. Modern algorithms are good: with the phone in a pocket or bag, counts are typically within five to ten percent of a research-grade device on flat ground. The known failure modes are predictable. It undercounts when the phone is not on you — on the desk, in the other room — which is the largest source of error and why phone counts usually run below wearable counts. It undercounts slow shuffling and pushing a pram or trolley, where the arm and body do not swing normally. It overcounts on bumpy vehicle rides and some cycling. Aura's walk tracker uses the phone motion sensor for steps and GPS for route, distance and pace during a tracked walk, which is the setup that gives the most reliable numbers: a deliberate walk with the phone on you, rather than passive all-day counting. For all-day totals, treat the phone number as a floor and the trend as the signal.

Steps are the friendliest number and the least informative, because a step is not a fixed unit — stride length varies with height, speed and terrain, and 8,000 slow steps and 8,000 brisk ones are different doses. Distance is better for comparing walks over time. Pace — minutes per kilometre — is the number that captures intensity, and the research is fairly clear that intensity matters: a cadence around 100 steps per minute, which is a brisk walk for most adults, is where the cardiovascular benefit per step is highest. A useful way to read a tracked walk is: distance tells you what you did, pace tells you how hard, and steps tell you how it compares to your daily average. Aura shows all three live on the map during a walk and saves them with the route. Over a month, the number worth watching is the seven-day average step count — daily counts swing too much with weather and schedule to mean anything on their own — and the second is the average pace of your deliberate walks, which should be creeping down as you get fitter.

Raise the seven-day average by about a thousand steps every two weeks. That is roughly ten minutes of additional walking a day, and it is small enough that nothing else in the day has to change to accommodate it. Weeks one and two: one deliberate ten-minute walk a day at a fixed trigger — after lunch, after the school run, the first thing after closing the laptop. Weeks three and four: make it fifteen, or add a second ten. Weeks five and six: one of the daily walks becomes brisk, at a pace where talking is possible but singing is not. By week eight, most people who started at 3,000 are averaging 6,500-7,500 without a gym, a membership or an hour they did not have. Two rules that make it hold. The walk is a habit, so it goes in the tracker with a streak and a floor version (five minutes, round the block) for the days it rains. And the reward is immediate: Aura credits XP and AuraCoins per thousand steps on a tracked walk, which is a small, closed feedback loop for a behaviour whose real reward is months away.

It does a great deal. The cohort evidence for walking is among the strongest for any single behaviour: lower cardiovascular and all-cause mortality, better blood-pressure and glucose control, and — with good evidence from randomised trials — meaningful improvements in mood and anxiety, which show up within a few weeks and which people notice before any physical change. It is also the most sustainable form of exercise there is, because it needs nothing and can be attached to things you already do. What it does not do, on its own: build significant muscle or bone density, which needs resistance work, especially past forty; produce large weight loss without dietary change, because the energy cost of walking is modest; or replace the higher-intensity work that improves cardiorespiratory fitness fastest. The sensible framing is that a walking base is the floor of an active life, not the whole building — and that for someone currently sedentary, the floor is where almost all of the available benefit is.

First, do not let the count become the walk. A step target hit by pacing the kitchen at 11pm is a target, not a health behaviour; the benefit comes from sustained walking at a decent pace, and the count is a proxy for that. Second, watch the average and ignore the days. A seven-day average is stable enough to plan against; a single day is weather. Third, keep the phone number honest by knowing its errors: a phone left on the desk for the afternoon produces a low day that is a measurement artefact, not a lazy day, and a tracked walk with the phone on you is the most reliable number you will get. Aura's walk tracker is built around that: start the walk, allow motion and location with one tap, walk with the app open, stop and save. The route, distance, pace and steps are stored under your account and go nowhere else. Use them to see the trend; let the trend, not the target, decide whether the plan is working.