How to measure body fat: how accurate scales, bioimpedance and scans really are

There are four ways to measure body fat that you can actually get outside a lab: a smart scale at home, bioimpedance analysis on a professional device, skinfold calipers, and a low-dose X-ray body composition scan. None of them measures fat directly. Each one measures something easier and converts it with a formula.

For a single reading the scan is usually the most reliable of these, and for a trend over months a scale or calipers can do the job, as long as you measure under the same conditions every time.

Why every method gives you a different number

You can’t put the fat of a living person on a scale. So every device measures a proxy (electrical resistance, the thickness of a fold of skin, how much an X-ray beam weakens on its way through you) and runs it through an equation built on a group of people who aren’t you.

Researchers judge all of these against the four-compartment model, which combines several lab measurements of body volume, body water and bone mineral. You won’t find one in a gym.

Almost every week someone shows me two numbers, one from their bathroom scale and one from the machine at their old gym, several points apart. Neither device is broken. They just use two different formulas.

Smart scales: accurate for weight, rough for fat

A smart scale uses bioimpedance. A weak current runs from one foot to the other, the scale estimates your body water from the resistance, and your fat from that. The current only travels through your legs and pelvis. Your upper body is filled in by software. I don’t mind people owning one. I mind them doubting a whole plan because of Tuesday’s reading.

The Swiss consumer programme Kassensturz had ten of the country’s best-selling body fat scales lab-tested in early 2025, priced from CHF 19.95 to 89.95. For fat they were 3 to 6 percent off the reference device, for muscle up to 28 percent, and six of the ten were rated “insufficient” (SRF Kassensturz, 2025). The most expensive scale did worst.

The finding I find more useful comes from a study of 15 bioimpedance devices, 14 of them consumer models. Measured twice in a row, they gave almost exactly the same number. Against the four-compartment model, though, they ranged from 3.5 points too low to 11.7 points too high (Siedler et al., 2023). In other words, your scale can show you the same wrong number every morning, very reliably. That still makes it usable for a trend: in the same study, some eight-electrode devices and a few foot-to-foot scales tracked change over 12 to 16 weeks reasonably well.

My rule for a scale at home: morning, fasted, after the toilet, barefoot, before training. And never read one day. Read the weekly average.

Bioimpedance analysis on a professional device

The bioimpedance analysis you get at a gym or a pharmacy works on the same principle, with electrodes for hands and feet and usually several frequencies. It reads arms, legs and trunk separately and prints a long report. Long is not the same as accurate. I read those printouts for the direction, not for the decimal.

A 2026 systematic review pooled twelve studies comparing bioimpedance with the four-compartment model in healthy adults. On average the error for body fat was small, between minus 3.5 and plus 4.4 points. For the individual, the limits of agreement were typically 15 to 20 points wide, and the authors concluded that bioimpedance is not equivalent to the reference at the individual level (Oliver et al., 2026).

Then there’s water. In one trial, 20 adults drank 11 ml of water per kilo of body weight after an overnight fast, so about 770 ml for someone weighing 70 kg. The professional analyser booked all of that extra weight as fat: an hour later it showed roughly 1.3 kg more fat mass (Tinsley et al., 2022). To be fair, another study with three simpler devices found only small changes after eating, drinking or exercise (Randhawa et al., 2021). The evidence isn’t settled. I still measure fasted every time, because it costs nothing.

Skinfold calipers: cheap, but only as good as the hands that use them

Calipers measure the thickness of a skinfold at several sites, in millimetres, and that part is precise. The trouble starts when you convert millimetres into a percentage. There are over a hundred equations for it, and the same measurements from one footballer gave anywhere between 4 and 8% body fat depending on the equation (Kasper et al., 2021). Pinch one centimetre away from the right spot and the result moves. Fifteen years of doing this taught me that the hands matter more than the caliper.

The same review still comes out in favour of skinfolds, because they’re the method least affected by a meal, a workout or your hydration. So I’d rather you write down the sum of the millimetres and compare it with last time than translate it into a percentage. And the same person takes the measurements every time. Change the person holding the calipers and the comparison is worthless.

The body composition scan

The scan reads, at every point of your body, how much two weak X-ray beams are weakened, and separates fat mass from lean mass. That means it sees where the fat sits: arms, legs, trunk, left and right. It estimates visceral fat, the fat around your organs, from the image. A scale or caliper can at best estimate it. On the printout I look at lean mass first, not the body fat percentage at the top of the page, because lean mass is what tells me whether the training is working.

It isn’t perfect either. In the Kasper review, eating a meal before the scan raised measured fat mass by up to 2.6%, and full glycogen stores in the muscles raised lean mass by 2.5% (Kasper et al., 2021). That’s why at VAULT Personal Training, Brandschenkestrasse 150, 8002 Zürich (Enge), we ask for two hours without a big meal before the scan, and we scan before a session, not after one. How the scan works and how repeatable it is from one scan to the next, I’ve explained in detail in What is a body composition scan.

The scan uses ionising radiation, at a very low dose. Under the Swiss Radiation Protection Ordinance every scan needs an individual justification, and we don’t offer it for lifestyle or wellness tracking. With us it’s optional and booked separately. When there’s a reason for it, we use it as a baseline before a training block and as a check afterwards. What it shows and how the appointment runs is on the body composition scan page.

Measuring body fat: the methods side by side

MethodWhat it measuresWeak pointWhat I use it for
Smart scaleResistance foot to footFills in the upper body, a single reading is often several points offWeekly average at home
Bioimpedance analysis (professional device)Resistance in arms, legs and trunkWide spread for the individual, sensitive to water and foodTrend, same device, always fasted
Skinfold calipersSkinfold thickness in mmThe percentage depends on the equation and on who measuresSum of millimetres, same person every time
Body composition scanFat and lean mass by regionRadiation, needs a justification, also affected by foodBaseline and check at the end of a training block

What I’d use, depending on what you want to know

If you only want to know whether you’re heading the right way, a scale, a tape measure round your waist and the weights you lift are enough. When your waist shrinks and the bar gets heavier, you’re very likely losing mostly fat and little muscle. You don’t need an expensive device for that. In my experience the tape measure is the most underrated tool on this whole list.

If you want to know whether you’re losing muscle while you diet, scales and calipers are too coarse. That’s exactly where you need a measurement that separates fat from lean mass. Which range is normal for your age is in healthy body fat percentage for women and men.

And sometimes fat is the wrong question. If what you really want to know is how fit you are, a VO2max test answers that far more directly than any body fat number.

Whatever you measure with, the number on its own changes nothing. At VAULT every set is written down, so next to the scan we can also see whether your strength goes up while the fat comes down. If you want a personal trainer in Zürich who keeps an eye on both, start with a free trial session.

Frequently asked questions

Are body fat scales accurate?

For weight, yes. For body fat, only roughly: in the Kassensturz test, ten scales were 3 to 6 percent off the reference for fat and up to 28 percent off for muscle (SRF Kassensturz, 2025). They’re useful for a trend if you always weigh at the same time and fasted.

What is the most accurate way to measure body fat?

The research reference is the four-compartment model, which only exists in labs. Of what you can get outside a lab, a body composition scan usually gives the most reliable single reading and separates fat and lean mass by body region straight from the image. Bioimpedance can be far off for an individual (Oliver et al., 2026).

When is the best time to measure body fat?

In the morning, fasted, after the toilet and before training. Water and food can move the number: in one trial a bioimpedance device showed about 1.3 kg more fat after roughly 770 ml of water (Tinsley et al., 2022). Keeping the conditions identical every time matters more than the exact hour.

How often should you measure body fat?

With a scale at home you can measure daily, as long as you only read the weekly average. When there’s a reason to repeat a scan, we do it after a few months at the earliest, for example at the end of a 12-week training block. Before that, smaller changes in lean mass barely stand out from the measurement error. Measuring more often mostly gives you noise.

If you want to work on the number as well, book your free trial session and we’ll work out together which measurement makes sense for your goal.