Why Your Devices Disagree About Your Weight and Your Steps

You step on the scale at home, then on the one at the gym, and they’re a kilogram apart. Your phone says 9,200 steps and your watch says 7,800. Neither device is necessarily broken. They measure differently, at different times, in different places, and knowing why helps you decide what to track.

Weight: Two Scales, Two Numbers

Your Weight Really Does Change During the Day

Your body weight moves with what you’ve eaten and drunk, what you’ve passed, and how much water you’re holding, which depends on salt, carbohydrates, exercise, and heat. Weighing yourself in the morning and again in the evening will rarely give the same number, so two scales used at different times of day are measuring two different weights.

The Scale Itself

• Calibration and quality. Bathroom scales vary in how accurate they are, and their error can differ across the range: one may read close at 60 kg and drift at 90 kg.
• The floor. Most scales need a hard, level floor. On carpet or an uneven tile, the same scale can read differently.
• Moving it. Many digital scales re-zero themselves when moved. Stepping on and off once after moving it, before the real reading, can help; check its manual.
• Clothes and shoes. Obvious, but they add up, especially at the gym.

Body Fat Is Harder Still

Smart scales estimate body fat with bioelectrical impedance: a weak current runs through your feet, and the scale estimates your body composition from how easily it passes. Water conducts well and fat poorly, so the estimate depends on how hydrated you are. A US National Institutes of Health panel that assessed the method stressed measuring under standard conditions, because eating, drinking, exercise, and hydration all change the result, and the European Society for Clinical Nutrition and Metabolism’s review of it says much the same: results are most reliable under standardized conditions, with equations validated for people like the one being measured (see the sources below). A home scale’s single morning reading is a long way from those conditions.

Two scales also use different equations for the same measurement, and they may report different quantities under similar names. Kaizora keeps these apart on purpose: Withings’ “muscle mass”, for example, includes smooth muscle and its water, so it’s kept as total muscle mass, not skeletal muscle mass.

Which to Trust

For tracking change over time, consistency beats accuracy: one scale, on the same hard floor, at the same time of day, under the same conditions (most people choose first thing in the morning, after the bathroom, before eating). A scale that’s a kilogram off but consistent still shows your trend correctly. Look at the weekly average rather than any one morning.

Steps: Phone and Watch

Where They’re Worn

A phone in your pocket and a watch on your wrist feel different motions. A watch can miss steps when your arm is still, such as pushing a stroller or carrying bags, and can count arm movements that aren’t steps, such as folding laundry. A phone left on the desk counts nothing while you walk around the office.

Different Algorithms

Each device turns movement into steps with its own method, tuned differently. In one well-known study, people walked a counted number of steps on a treadmill wearing several devices: the smartphone apps came within about 7% of the true count, while the wrist-worn trackers ranged from close to undercounting by more than a fifth (Case and colleagues, JAMA, 2015). Devices have improved since, but the point stands: two methods give two counts.

Counting the Same Walk Twice

If both devices send their steps to one place, a naive total would count your walk twice. Apple Health, for example, picks one source at a time when it totals your steps, by an order of priority you can change in the Health app.

Which to Trust

Again: choose one and stick with it for trends. A watch you wear all day usually sees more of your day than a phone you put down; a phone you always carry can be the better choice if you don’t wear a watch at night or at your desk.

How Kaizora Handles It

• Every value names its device. When a value comes from an import, Kaizora records the device that measured it, and you can choose a device when you log one yourself. On a metric’s page, filter by device to see each one’s numbers on their own. See Add Your Devices.
• Totals count a walk once. For a metric that adds up, like steps, a day’s total takes what you logged yourself plus the one imported service with the largest total, and within that service, the device with the largest total. A phone and a watch counting the same walk don’t add up to twice the steps.
• Averages show both. For a metric that’s averaged, like weight, readings from two scales both count, so filtering by device is the way to compare them.
• Exact definitions. Each metric is one precisely defined quantity, so a vendor’s estimate isn’t merged with a different one under a similar name.

The Short Version

• Your weight changes through the day; two scales at two times measure two weights.
• Body fat from a smart scale depends a lot on hydration and timing.
• Phones and watches feel different movements and count them differently.
• For trends, use one device, the same way every time, and look at averages.

Sources

• National Institutes of Health, “Bioelectrical impedance analysis in body composition measurement: NIH Technology Assessment Conference Statement,” American Journal of Clinical Nutrition 64, no. 3 (1996): 524S–532S. doi:10.1093/ajcn/64.3.524S
• Kyle UG et al., “Bioelectrical impedance analysis—part I: review of principles and methods,” Clinical Nutrition 23, no. 5 (2004): 1226–1243. doi:10.1016/j.clnu.2004.06.004
• Case MA, Burwick HA, Volpp KG, Patel MS, “Accuracy of smartphone applications and wearable devices for tracking physical activity data,” JAMA 313, no. 6 (2015): 625–626. doi:10.1001/jama.2014.17841