Body fat percentage measurement methods: DEXA vs BIA vs skinfolds vs photos

Every body fat measurement method has an error margin, and no method gives you your "true" number, because outside of a cadaver lab that number does not exist as something you can read off a device. What you can read is an estimate, and every estimate has a band of uncertainty around it: plus or minus a percentage point or two for the good methods, plus or minus ten for the bad ones. Knowing the size of that band for the method you use is far more useful than hunting for a single trustworthy figure, because the goal of measuring body fat is not accuracy to 0.1 percent. It is consistency: measuring the same way under the same conditions so that the trend over months tells you whether you are gaining muscle, losing fat, or standing still. This guide ranks the methods by how accurate AND how practical they are, tells you honestly where each one lies to you, and gives you a workflow that turns noisy weekly numbers into a trend line you can actually trust.

The methods, ranked by accuracy and practicality

Accuracy alone is a bad way to choose. The gold-standard lab methods are accurate and almost nobody uses them monthly, because a $100 scan you take twice a year tells you nothing about whether last month's cut worked. The table below ranks each method on both axes: how tight its error margin is, and how realistic it is to repeat often enough to build a trend.

Method	Error margin	Cost	Practicality
DEXA scan	±1-2%	$50-150 / scan	Lab only, gold standard
BodPod (air displacement)	±2-3%	$50-100 / session	Facility only
Hydrostatic weighing	±2-3%	$40-75 / session	Hard to find, awkward dunk
3D body scanner (Naked, Styku, Fit3D)	±3%	$1-3K device or ~$30/scan	Increasingly common at gyms
Skinfold calipers (trained tester)	±3-5%	$20-50 one-time	Needs technique and a tester
Calipers, DIY no training	±10%+	$20-50 one-time	Unreliable, skip it
BIA (body comp scales, handhelds)	±5-10%	$30-150 one-time	Daily at home, very convenient
Photos + waist measurement	No number	Free	Excellent trend signal

Read that table twice, because the ranking is not top-to-bottom good-to-bad. DEXA is the most accurate single reading you can buy, but photos plus a tape measure, which produce no body fat number at all, often beat a BIA scale for tracking real progress over a cut, because they are not fooled by how much water you are holding. Accuracy of one reading and usefulness of a trend are two different things.

The honest reality: BIA scales lie

The smart scale in your bathroom that flashes a body fat percentage is running bioelectrical impedance analysis. It sends a tiny current up one leg and down the other and times how long it takes, because fat and lean tissue conduct electricity differently. The problem is that the current does not actually measure fat. It measures conductivity, and conductivity is dominated by how much water is in your body and where that water is sitting right now.

That makes the reading hostage to your hydration, the time of day, what you last ate, whether you trained, and even skin temperature. The same scale can show you five different body fat numbers across a single day on the same body: lower in the morning after an overnight fast and mild dehydration, higher in the evening after meals and fluids, different again right after a sweaty session. None of those swings reflect a real change in fat. The fat on your body does not move 5 percent between breakfast and dinner. The water does, and the scale cannot tell the difference. A foot-only scale is the worst of the lot because it only measures your lower body and extrapolates the rest. This is not a reason to throw the scale out. It is a reason to never trust a single reading from it, and to only ever use it as one input to a trend taken under fixed conditions.

The consistency rule

Here is the rule that makes any method useful, even a mediocre one: pick ONE method and take every measurement under identical conditions. Same time of day, same hydration state, same point relative to meals, same training state, repeated every 4 to 6 weeks. The absolute number the method spits out barely matters. The direction it moves, measured the same way each time, is the whole signal.

This works because a consistent method has a consistent error. If your BIA scale reads 3 percent high every single time you step on it first thing in the morning, fasted, before training, then it reads 3 percent high in January and 3 percent high in March, and the difference between those two readings cancels the error out. You lose the accuracy of the absolute figure but you keep the accuracy of the change, and the change is what tells you the cut is working. The fastest way to make any method worthless is to vary the conditions: weigh fasted one month and after lunch the next, and you have buried a real 2 percent drop under a 4 percent measurement swing.

What to use when

The right method depends on what you are tracking and what you can afford to repeat.

• Tracking long-term recomp progress: DEXA every 3 to 6 months for the accurate anchor, plus monthly photos in between to catch the trend the scans are too infrequent to show. This is the setup if you are running a slow body recomposition where the scale barely moves but your shape is changing.
• Cutting or bulking phase tracking: skinfolds every 4 weeks with the same tester, OR a BIA scale plus waist measurement weekly under fixed conditions. During a bulk the waist tape is your early-warning system: the lean bulk surplus guide uses waist creep as the stop-rule that tells you the surplus has tipped from muscle into fat.
• No budget for DEXA: BIA plus waist plus photos is the budget stack, and the triangulation is what makes it useful. Any one of the three is weak on its own. Together they corroborate each other: when the scale weight drops, the waist shrinks, and the photos show more definition, you do not need a lab to know fat is leaving.
• Race prep or contest prep: weekly skinfolds with the same coach, every week, same sites, same hands. At the sharp end of a cut you need a high-frequency read and a fixed tester is the most repeatable thing you can get without a lab on call.

Common mistakes

1. Comparing your BIA scale's number to a friend's DEXA. These are different methods with different error directions. Your scale might read high and their DEXA low, so a "match" or a "gap" between the two numbers means nothing. You can only compare a method to itself.
2. Re-measuring weekly with BIA expecting weekly truth. A good week of cutting moves your real body fat by a few tenths of a percent. Your BIA reading swings by several percent on water alone. Week to week, the noise is bigger than the signal, so weekly BIA readings look like random static. Take them weekly if you want, but only read the 8-week trend, never the week-on-week jump.
3. Trusting morning BIA when you are cutting. A cut depletes muscle glycogen, and glycogen binds water. Less stored glycogen means less water in the muscle, which BIA reads as less lean tissue, which it reports as lower body fat. So a cut can make your morning BIA number drop faster than your actual fat, flattering you into thinking the diet is working better than it is. Pair it with the waist tape, which is not fooled by glycogen.
4. Skipping photos because "I look the same daily." Of course you do. Fat loss at a healthy rate is invisible at the day-to-day scale, which is exactly why you compare month-over-month photos, not day-over-day mirror checks. The change between this month's photo and the one from eight weeks ago is obvious in a way that staring at yourself every morning will never show you.

Visual benchmarks: rough body fat by appearance

If you have no device at all, your reflection is a usable rough gauge. These ranges are approximate, vary with how much muscle you carry, and are best used to sanity-check a number rather than replace it. Men and women store fat differently, so the ranges are not interchangeable.

Men:

• 6 to 9%: visible vascularity, clear separation between muscle groups, stage-ready and not maintainable year-round.
• 10 to 12%: visible six-pack, athletic look, the lean end most lifters can hold.
• 13 to 17%: a blurry four-pack, healthy and sustainable.
• 18 to 22%: no visible abs, a softer midsection.
• 23 to 30%+: a visible belly, the "skinny fat" or overweight range.

Women:

• 16 to 19%: visible vascularity in the arms, athletic and at the low, hard-to-hold end.
• 20 to 23%: visible muscle definition, a fit look.
• 24 to 28%: athletic but softer.
• 29 to 33%: the average, healthy range.
• 34%+: above-average storage.

The Body Comp Tracker workflow

Put the rules together into a routine you can actually keep. Log your weight, your BIA reading, and your waist measurement weekly, all taken under the same conditions: same morning, fasted, after using the bathroom, before training. Take front and side photos once a month in the same lighting and pose. Then ignore the individual weeks. Over 8 to 12 weeks the trend line across all three measures is the truth, and any single week is noise you should not react to.

This is exactly what the Body Composition Tracker is built for: it holds weight, BIA, waist, and photos together so the trend is visible at a glance instead of buried in a notes app, and the AI coach reads the trend rather than the latest panicky reading. Body fat percentage is a decision input, not a vanity metric. It tells you whether to keep cutting or to start a maintenance phase, and many of those decisions key off your current level: the cutting calories guide uses your starting body fat to set the deficit, and your protein target on a recomp is set per kilo of bodyweight that your measurements help you track. A noisy number you measure consistently drives those decisions perfectly well. A precise number you measure carelessly drives them off a cliff.

Frequently asked questions

What is the most accurate way to measure body fat percentage?

DEXA is the most accurate method you can realistically book, with an error margin of about ±1 to 2 percent, which is why it is treated as the gold standard for body composition. BodPod and hydrostatic weighing are close behind at ±2 to 3 percent. But the most accurate single reading is not automatically the most useful for tracking: a DEXA scan twice a year cannot tell you whether this month's cut worked, while a cheaper method measured consistently every few weeks can. For most people the best practical choice is a DEXA anchor a couple of times a year plus a consistent home method in between.

How accurate are body fat scales?

Body fat scales use bioelectrical impedance and carry a wide error margin of roughly ±5 to 10 percent, and worse, the reading swings with your hydration, the time of day, your last meal, and whether you just trained. The same scale can show five different numbers across one day on the same body. Treat the absolute number as unreliable and only ever read the long-term trend, always measuring at the same time of day under the same conditions so the error stays consistent.

DEXA scan vs bioelectrical impedance, which should I use?

Use both for different jobs. DEXA gives you an accurate absolute number a few times a year as a reliable anchor. BIA gives you a cheap, daily-at-home reading whose absolute value you should distrust but whose trend, measured under fixed conditions, is useful between scans. The mistake is comparing a BIA number directly to a DEXA number: they are different methods with different error directions, so they will rarely agree, and the gap tells you nothing. Compare each method only to itself over time.

Why does my body fat percentage change every day?

It does not. Your actual fat does not move by several percent between morning and evening. What changes is your body water, which dominates a BIA reading: hydration, food, sodium, and glycogen all shift water around and the scale misreads those shifts as fat changes. On a cut this gets worse, because depleting muscle glycogen pulls water out of the muscle and BIA reports that as a fat drop that has not really happened. Pin your conditions down and only trust the multi-week trend.

Can I track fat loss without measuring body fat percentage at all?

Yes, and for many people it is the better approach. Photos plus a waist measurement give you no number but an excellent trend signal, and they are free and not fooled by hydration. Take monthly front and side photos in the same lighting and pose, and measure your waist weekly under fixed conditions. When the waist shrinks and the month-over-month photos show more definition, you are losing fat, no body fat percentage required. The number is a decision input, not the goal itself.

Pick one method, hold the conditions steady, and read the trend, not the reading. DEXA for the accurate anchor, BIA plus waist plus photos for the cheap high-frequency picture in between, and your reflection as a rough sanity check. The Body Composition Tracker keeps all of it in one trend line so the eight-week direction is obvious, and the body recomposition guide is the right next read if your scale weight is stubbornly flat while your shape keeps changing.