Body Composition Analysis: What the Numbers Actually Mean
By Paul Butler · Online Coach & Personal TrainerYour bathroom scale is a miserable so-and-so, and it is also lying to you. Not by much. But by enough to ruin a Thursday.
Body composition analysis is the attempt to fix that: instead of one number for everything, it splits you into fat, muscle and water, usually by way of a machine in the gym that you stand on barefoot while it hums. The numbers it gives you look very precise. Whether they are accurate is a different question, and it has been tested properly. So here is what body composition analysis actually measures, how far out the machines are, and which numbers are worth your attention.
What does body composition analysis actually measure?
It measures how much of your weight is fat and how much is everything else: muscle, bone, organs and water. The reference method is a DXA scan, a low-dose X-ray that costs money and lives in clinics. The gym version, usually an InBody or similar, passes a small current through you and estimates the split from how easily it travels. Fat resists, water conducts, and the machine does the maths.
| Method | What it measures | How far out | Good for |
|---|---|---|---|
| Bathroom scale | Total weight only | Accurate for weight; swings a kilo or two a day with water and food | A weekly average, never a single day |
| Gym body composition machine | Estimates fat, muscle and water from an electrical signal | Body fat read 3 to 4 points low on average versus DXA; very repeatable | Trend over months, same conditions each time |
| Tape measure | Waist, hips, arms, thighs | Only as consistent as where you put the tape | Cheap, honest, and waist tracks fat loss well |
| DXA scan | Fat, lean mass and bone, by region | The reference the others are judged against | A snapshot, if you want to pay for one |
The middle row is the one people obsess over, so it is worth knowing where that "3 to 4 points low" comes from.
How accurate is the machine in your gym?
The best real-world test I know of put a thousand healthy, active adults through an InBody and a DXA scan in ordinary conditions rather than a fasted, rested laboratory. The machine tracked the scan closely across the group, but with a consistent lean in one direction: it underestimated fat and overestimated muscle, by about four percentage points in men and three in women.
"In uncontrolled conditions, MFBIA provides moderately accurate method for measuring TBW and body composition (except for VAT)."
MFBIA is the machine, TBW is total body water, and VAT is visceral fat, the number the printout gives you for fat around your organs, which the study found was the least reliable figure on the sheet. Two honest caveats: those thousand people were fit young military personnel, and the error is generally thought to be larger the more body fat someone carries. So if you are starting from a heavier place, treat the absolute number with even more suspicion.
The good news in the same study is that the machine was very consistent with itself. Stand on it twice under the same conditions and you get the same answer. That is exactly what you want from a trend tool. Ten weeks apart, same time of day, same hydration, the change is real even if the starting number is not.
Why your body fat went up after you started training
This is the message I get most, usually with a screenshot and some alarm. Someone starts lifting, gets serious about water, maybe starts creatine, and the machine reports more fat two weeks later. Nothing has gone wrong. The machine reads through water, and their water has changed. Muscles holding more fluid, a bigger breakfast, a session an hour earlier: all of it moves the estimate by more than two weeks of actual fat loss ever could.
The machine does not care that you had a good fortnight, any more than the bathroom scale cares that you had a bad week. Neither is judging you. They are just measuring the wrong thing on the wrong day.
The numbers actually worth tracking
In order of usefulness for someone trying to lose fat and get stronger, not compete on a stage.
- A weekly average weight. Weigh daily if you can stomach it, average the seven, compare the averages. Individual days are noise.
- Waist, at the navel, same tape, same spot. The cheapest honest measure of fat loss there is.
- What you lift. If the weights are going up and the waist is coming down, the composition is changing in the right direction, whatever any machine says.
- Photos, monthly, same light. Painful to take and the most convincing evidence you will ever have.
- The gym machine, monthly at most, under identical conditions, read for direction only.
If you are in a calorie deficit and the scale has stalled while your clothes fit better, this is usually why, and it is a measurement problem, not a metabolism problem. And if you are hoping to lose fat and build muscle at the same time, the scale may not move at all, which is the one situation where composition genuinely matters more than weight.
When the number is the problem
Thirty day transformations sell a lie about how bodies change, and a machine that prints a body fat percentage to one decimal place quietly sells the same lie: that change is fast, precise and visible on demand. It is not. It is slow, noisy and mostly invisible week to week, which is why people who measure daily tend to quit and people who measure monthly tend to finish.
If stepping on a scale has started to decide your mood for the day, stop stepping on it. That is not a small thing, and it has a name and a cause. If it goes further than mood, and food or weight is taking up more of your head than feels right, that is a conversation for your GP, not a blog. This is general information, not medical advice.
Frequently asked questions
How often should I do a body composition scan?
Monthly at most, and only if the conditions are the same each time: same machine, same time of day, before food and training, similar hydration. Weekly scans just measure your water and your breakfast.
Are the body fat scales you buy for home any good?
They use the same principle as the gym machine, usually through the feet only, so they estimate from your legs and guess the rest. Treat the fat percentage as a rough trend and the weight as the useful bit. Do not compare their number to anyone else's.
Does drinking water before a body composition test change the result?
Yes, noticeably, because the machine reads through water. A big drink, a big meal, a hot day or a training session beforehand will all shift the estimate. Whatever you did before the first scan, do the same before every scan after it.
Is body composition analysis worth doing at all?
For most people, a tape measure and a weekly average weight tell you everything the machine does, for nothing, with less room to panic. Where the machine earns its place is when weight is not moving and you want confirmation that the composition is, as long as you read it for direction rather than decimals.
DM me STARTING on Instagram @paulbutler.coach and I will send you The Starting Point: five days of the fundamentals, then a four week plan built round your real week, with the measurements that matter and none that do not. No cost, no catch.
If you are in Southampton and want it done properly in the room, the face to face coaching at JD Gyms Bitterne is on the Southampton page, price included. Or book a free 15-minute call and we will work out what is worth measuring for you.
This article is general information, not medical advice. If food or weight is taking up more of your head than feels right, or you have a history of disordered eating, please talk to your GP.
More honest guides on fat loss, nutrition and training
The accuracy figures are Potter et al. 2025, European Journal of Clinical Nutrition 79(12):1235-1244, 1,000 adults measured on an InBody 770 against DXA, quoted verbatim and linked in the text. Images from Pexels (free to use): SHVETS production, Mikhail Nilov, cottonbro studio.
