How your phone counts your steps (and why the number never quite agrees)

You walk, the number goes up. Sounds simple. But behind that counter sits a sensor, an algorithm and a few traps, which is why two devices never tell you exactly the same thing.

The sensor: an accelerometer

Every smartphone has an accelerometer inside, a tiny sensor that measures movement along three axes. When you walk, your body moves in a fairly steady rhythm: a small jolt each time your foot lands, a gentle sway back and forth.

Your phone looks for that pattern. When it spots a run of movements that looks like walking, it counts steps. When the movement is erratic (you fiddling with the phone, a car vibrating), it’s supposed to ignore it.

On most recent phones, this job goes to a dedicated low-power chip. That’s why step counting runs all day long without draining your battery.

Where your steps are stored

Your phone doesn’t keep those steps to itself. It files them in one central health hub.

• On iPhone, that’s the Health app. Apps you’ve approved can read your steps from it.
• On Android, it’s Health Connect, the system’s health vault. Same idea: you decide which apps can read which data, and you can revoke access at any time.

So an app like Kiwalk doesn’t count your steps itself. It reads the total your phone has already recorded. The upside: no need to keep the app open, and no extra battery drain.

Why the number varies

1. Your phone isn’t always on you

This is the big one, by a long way. A study published in 2019 in JMIR mHealth and uHealth spent a week comparing the steps counted by the iPhones of 54 adults with those from a pedometer worn on the waist. The result: the iPhone counted on average 12% fewer, roughly 1,300 steps a day. The gap was wider for people who often carried their phone in a bag or set it down.

The researchers put it plainly: when the phone isn’t on you, it can’t count.

2. Where you carry it

In a trouser pocket, your phone follows the rhythm of your legs nicely. In a bag that swings around, or in a hand that’s also holding an umbrella, the signal gets muddier.

3. How you walk

A very slow stroll, shuffling around the grocery store, pushing a cart or a stroller: all situations where the rhythm is harder to recognize.

4. Double counting across devices

If you wear a watch and carry a phone, both are counting. The Health app and Health Connect have rules to avoid counting twice, but depending on your settings, the total you see may come from one or the other.

How to make your step counter more reliable

• Keep your phone on you, ideally in a pocket, whenever you’re on the move.
• Check permissions: in Health Connect (Android) or the Health app (iPhone), read access to steps needs to be turned on for the app that uses them.
• On Android, keep an eye on battery optimization. Some manufacturers clamp down hard on background apps. If your steps aren’t syncing properly, let the app run without restrictions.
• Pick one main source if you have a watch, so the devices don’t surprise you.

What about Kiwalk?

Kiwalk reads your steps and distance from Health Connect or the Health app, nothing else and no GPS. We check that they really came from your own two legs, then every tier you reach earns you Kiwis. If your count looks low, the tips above fix the vast majority of cases.

See how Kiwalk counts and checks your steps

Frequently asked questions

Does my phone count steps when the app is closed?

Yes. Your phone’s system does the counting in the background. An app like Kiwalk just reads that total when you open it.

Why don’t my watch and my phone show the same number?

They aren’t worn in the same place or at the same times. Your watch goes everywhere with you, while your phone sometimes stays on the table. Each one also runs its own algorithm.

How do I make my step counter more accurate?

Keep your phone on you, ideally in a pocket. Check that step access is allowed, and on Android, make sure battery optimization isn’t blocking the app.

Sources

1. Amagasa S. et al., “How Well iPhones Measure Steps in Free-Living Conditions: Cross-Sectional Validation Study”, JMIR mHealth and uHealth, 2019. mhealth.jmir.org