Average Steps per Day by Age, Sex and Country: Benchmarks for Health Apps and Programmes

The United States and Thailand walk almost exactly the same amount. In one of the largest studies of everyday walking, Americans averaged 4,774 steps a day and Thais 4,764. Yet among the people measured, obesity was 27.7% in the United States and 9.1% in Thailand [1].

The difference was not how much people walked on average. It was how unevenly. The study, by Stanford researchers using phone data from 717,527 people in 111 countries, found that the gap between a country’s most and least active people predicted obesity better than its average step count did [1]. That finding matters to anyone who sets step goals, benchmarks users or designs an activity programme, because almost all of them are built on averages.

This is what the data says about how much people walk, by country, sex and age, drawn from the study’s own published data [2], and why the average is the wrong number to build on.

How many steps does the average person take a day?

About 5,000. Across 68 million days of data, the global average was 4,961 steps a day [1]. The study measured steps with the iPhone’s built-in step counter through the Argus app, from people who recorded at least 10 days, and phones were carried for about 14 hours a day on average.

Two things push the true figure higher. Phones count only while they are in a pocket or bag, so steps around the house, at the gym or while the phone is on a desk are missed. And every step counter, phone or wrist, undercounts the short, slow bursts that make up much of real life; in free-living tests both miss roughly a quarter of steps, as we covered in step counting accuracy on phone and wrist. So read 5,000 as a floor measured the same way everywhere, which is what makes the comparisons below useful.

Which countries walk the most and the least?

From the study’s data for the 46 countries with at least 1,000 users [2]:

The United Kingdom averaged 5,444, Germany 5,205, Canada 4,819, the United States 4,774 and India 4,297.

Why is the average the wrong number?

Because two countries with the same average can have very different health. The Stanford team measured activity inequality: how unevenly steps are spread between the most and least active people, on a scale from 0, where everyone walks the same amount, to 1. Across the 46 countries, activity inequality tracked obesity much more closely than average steps did. We recomputed both from the published data: a correlation of 0.79 for inequality against −0.62 for the average [1][2].

Same-average pairs make the point. The United States and Thailand walk the same; inequality is 0.303 against 0.272. Australia and Taiwan both average about 5,000 steps; obesity among users was 21.4% against 6.6%, and inequality 0.304 against 0.262. The United Kingdom and Switzerland are within 70 steps of each other; obesity was 19.5% against 9.1% [2].

The researchers found that in the five countries with the widest gaps, people were 196% more likely to be obese than in the five with the narrowest [1]. The finding is a correlation across countries, and diet, income and culture differ too. But it says something specific: a population’s health depends less on how much it walks on average than on how many people are left near the bottom.

Two caveats belong next to these numbers. The obesity figures are for the study’s app users, from heights and weights they entered themselves, so they are not national rates. And the data comes from iPhone users in 2013 and 2014, a group that skews towards higher incomes and an interest in health [1]. Those biases apply to every country in the sample, which is why the comparisons between countries hold up better than any single country’s figure.

Do women walk less than men?

In all 46 countries, yes [1]. But the size of the gap is what predicts inequality. The gap between men and women was smallest in Sweden and largest in Qatar, and it accounted for 43% of the variation in activity inequality between countries. Where inequality was high, women’s activity fell further than men’s: moving from the lowest-inequality to the highest-inequality countries, median steps fell 39% for men and 58% for women [1].

That has a direct consequence for anyone designing a programme. A challenge that rewards the most steps, in a population where women already walk 20% to 30% less, mostly rewards men.

How many steps a day is normal for my age?

From the study’s data [2]:

A woman over 70 who takes 3,000 steps a day is at the bottom of the global distribution and exactly typical for her age and sex. Telling her she is “below average” is accurate and unhelpful.

Does where you live change how much you walk?

Yes, and more for women. Across 69 US cities, the researchers compared step counts with Walk Score, a 0 to 100 rating of how walkable a city is. A city 25 points more walkable was associated with 868 more steps a day for women and 622 more for men, and with lower activity inequality overall [1]. In the most walkable cities, steps peaked during the working day, at commuting times and lunch, rather than only in the evening.

Environment is not something an app can change, but it is something an app can account for. A user in a car-dependent suburb and a user in a dense city with the same step count are not making the same effort.

What should products and programmes do with this?

1. Benchmark people against people like them. Age and sex alone move the typical count by 3,000 steps a day, before device, country or environment. Compare a user with their own history first, and with people of similar age, sex and region second. A global average is the least informative comparison available.

2. Report the spread, not just the mean. A programme whose average rises from 5,000 to 5,500 might have moved everyone a little or moved the most active a lot. Report the share of people below 5,000 steps, or the bottom quarter’s average, alongside the mean.

3. Design for the least active. The researchers simulated adding 100 steps a day per person in two ways. Spreading it across everyone reduced obesity by a median of 1.0%. Concentrating it on the least active reduced it by a median of 4.0%, and by up to 8.3% [1][2]. The simulation assumes perfect uptake, which no programme achieves, but the direction is clear, and it matches the mortality evidence that gains are steepest at the low end of the step range, which we covered in do fitness trackers work.

4. Set goals from each person’s starting point. A fixed 10,000-step target is close to unreachable for the people who would gain most. Goals of a few thousand steps above a person’s own baseline work better, as we covered in the 10,000 steps myth.

5. Compare like with like over time. Steps drop about a tenth in winter, so seasonal comparisons need care, as we covered in why step counts drop in winter. Phone and wrist counts differ too, so do not mix them in one benchmark.

The average person takes about 5,000 steps a day. It is a useful fact for a headline and a poor foundation for a product. The same data shows what matters more: who is at the bottom of the distribution, how far they are from everyone else, and whether a programme reaches them at all.