You stepped on your smart scale Tuesday morning. It told you your body fat was 24.3 percent. You stepped on it again Friday morning. It said 21.1 percent. You did not lose fat in three days. You lost water, and the scale panicked.
Bioelectrical impedance analysis, or BIA, is the physics behind almost every consumer body-composition scale, and it has a real accuracy problem that no manufacturer explains clearly on the box. This piece unpacks what BIA actually measures, why the readings swing, when to trust them, when to ignore them, and the strict protocol that makes them useful anyway.
What BIA is actually doing under the glass
Bioelectrical impedance is exactly what it sounds like. A tiny alternating current, usually in the 5 to 500 kilohertz range, passes through your body. Consumer scales usually run a single frequency near 50 kilohertz. The device measures the resistance the body's tissues offer to that current.
Muscle, blood, and body water are excellent conductors of electricity. Fat is a poor conductor. Bone is essentially a nonconductor. By measuring total impedance and applying a manufacturer-specific regression equation that also incorporates your height, weight, age, and sex, the scale estimates fat-free mass and, by subtraction, fat mass.
Notice what the scale is measuring directly: the electrical properties of your body water compartment. It is not measuring fat. It is not weighing your fat cells. It is measuring how well a low-voltage current passes through you, and then guessing at fat from a formula that assumes your hydration is normal and your body-water distribution is average for your demographic. When either assumption fails, the number is wrong.
Why the reading swings three percent day to day
Independent validation studies routinely find that consumer foot-to-foot BIA scales carry a standard error of estimate of roughly 3 to 4 percentage points for body fat compared to the reference standard. In practice that shows up as day-to-day swings that have nothing to do with actual body-composition change.
The main confounders:
- Hydration status. This is the biggest. A dehydrated body has more concentrated body water, which raises impedance, which the scale interprets as more fat. A well-hydrated body reads leaner. The same person can swing 2 to 3 percentage points of "fat" across the same day.
- Recent meals and alcohol. Food and drink change body water and gut contents. Alcohol is diuretic and can register as dramatic overnight fat loss that reverses within 24 hours.
- Recent exercise. Blood pooling, sweat loss, and glycogen depletion all shift readings. A hard leg workout the day before will typically make the scale report lower body fat the next morning. That is not a real change.
- Skin condition. Cold, callused, or dirty feet increase contact resistance. Warm, moist feet decrease it. Both distort the reading in different directions.
- Menstrual cycle. Water retention in the luteal phase can add several pounds of body water and register as increased fat percentage. This is one of the most under-discussed sources of noise for female users.
- Ambient temperature. Cold peripheral circulation raises impedance. A summer morning and a winter morning are not producing directly comparable numbers.
None of these confounders reflect changes in your actual body composition. All of them show up as changes in the number on the scale.
Foot-only vs 8-electrode: the extra hands help
Most consumer smart scales are foot-to-foot. Current enters through one foot and exits the other. This measures impedance across your legs and lower trunk, and then extrapolates to the rest of your body using assumptions built into the formula. Your upper body is essentially estimated, not measured.
Higher-end scales, sometimes called 8-electrode, segmental, or InBody-style, add handles you hold. Current now enters and exits at four contact points, so the device can measure the trunk and upper body directly rather than assume them. Segmental analysis — separate estimates for each arm and leg — becomes possible.
Eight-electrode scales are more accurate for total body composition and much more accurate for anyone whose upper-body-to-lower-body ratio is unusual. Athletes with well-developed shoulders and moderate legs, or people carrying most of their fat centrally, get more reliable numbers from the segmental approach. Eight-electrode scales are also more expensive and slower to use. For most people wanting a home trend, foot-to-foot is fine as long as they understand the noise floor. For anyone whose training splits upper and lower body dramatically, the eight-electrode gap is real.
Multi-frequency BIA and the gym InBody scan
If you have ever stood on an InBody scanner at a physical therapy office or a commercial gym, you have used a higher-end version of the same technology. Those units run multiple frequencies simultaneously — typically 5, 50, 250, and 500 kilohertz — instead of the single 50 kHz used in most home scales. Because cell membranes act as capacitors, low frequencies flow around cells while high frequencies pass through them. Comparing the impedance at multiple frequencies lets the device distinguish intracellular from extracellular water, which is what a good clinical-grade unit is really doing.
Multi-frequency segmental BIA is closer to DEXA than a home foot-to-foot scale is, particularly for lean mass estimation and for tracking training-driven changes. It still is not DEXA. Peer-reviewed comparisons of InBody units against DEXA typically find total-body-fat agreement within 2 to 3 percentage points on average, with wider individual variance. That is a real upgrade over the single-frequency home scale, and it is the honest reason InBody scans have become a fixture in performance-oriented facilities.
For most people the practical protocol is to use a home BIA scale for weekly trend and to book an InBody or DEXA scan quarterly for a periodic calibration point. Home scale for cadence, higher-tier method for absolute numbers.
Why DEXA is the reference standard
When a validation study calls a method accurate, it is almost always compared against dual-energy X-ray absorptiometry, or DEXA. DEXA scans use two X-ray energies to distinguish bone, lean tissue, and fat tissue directly based on their differential absorption. The output is a three-compartment model — bone mineral, lean soft tissue, fat mass — with precision typically within 1 to 2 percentage points of body fat.
DEXA is available at many hospital imaging departments, sports-medicine clinics, and increasingly at standalone body-composition scan shops in major cities. A single scan usually runs 50 to 150 dollars in the United States. It is not something you do daily. It is something you do quarterly or twice a year to calibrate what your scale is telling you.
Air-displacement plethysmography (Bod Pod) and hydrostatic underwater weighing are the other reference-quality methods. Any of them will be tighter than any consumer BIA scale, at any price point, on any given morning.
The protocol that makes BIA useful anyway
None of the above means a consumer scale is useless. It means the way most people use one is useless. A daily-check-then-panic approach amplifies noise. A structured protocol collapses noise into signal.
- Same time of day, every measurement. First thing in the morning, before eating or drinking, after using the bathroom. Not "morning-ish." Exact.
- Same hydration state. This is the hardest one to control. The convention is to weigh after a full night of not eating or drinking, so you are consistently near the bottom of your daily hydration cycle. Do not weigh after a big glass of water. Do not weigh after a sauna.
- Same day of week. If you drink socially on Saturday, your Sunday morning weight will be dry-side and the reading will look lean. Compare Wednesday to Wednesday. Do not compare Wednesday to Sunday.
- Same feet, same standing position. Clean, room-temperature feet. Standing evenly. Feet in the same spot on the scale platform.
- Report averages, not individual readings. A single reading is noise. A four-week rolling average of six or eight measurements is signal. Do not react to any single day.
- Calibrate against DEXA once or twice a year. The absolute body-fat percentage from your scale is probably wrong by a couple of points either direction. The change in that number is what matters. Verify occasionally that the direction of change matches a reference method.
Under that protocol, a consumer BIA scale becomes what it is genuinely good at: a decent trend instrument. It will not tell you your true body-fat percentage. It will tell you, over months, whether you are losing fat or losing muscle when your scale weight drops, which is the question that actually matters for most people.
Why there is no smart scale in our catalog yet
Full transparency: we do not currently carry a body-composition scale. When we grade one on the Vyvata rubric, it will need documented independent validation studies against DEXA, transparent disclosure of the estimation equations used (which most manufacturers treat as proprietary), a reasonable safety profile, and something better than the standard 3 to 4 percentage-point error band on the accuracy dimension. Some models will clear the Verified bar. Most will not.
Rather than adding one before it is scored, we would rather point at what a beginner should actually do first: measure the things that are reliably measurable at home. Weight on a bathroom scale is accurate to about 0.1 pound. Waist circumference with a tape measure at the level of the navel is accurate to about a quarter inch. Both are strong indicators of the metabolic outcome you probably care about. Use those, plus periodic DEXA, rather than adding a noisy composition estimator on top.
The wrist-based alternative for trend seekers
If what you actually want is a trend line for your health across weeks and months, the ScanWatch 2 is a more informative device than a smart scale. Resting heart rate, overnight HRV, and sleep architecture respond to training, nutrition, and recovery on the same timescale that body composition actually changes on. Watching those numbers move quarter over quarter is a more honest read on whether the work you are doing is producing physiological change than watching a BIA scale swing by hydration.
None of that is a body-composition measurement. It is a different question with a cleaner instrument.
The honest bottom line
Bioelectrical impedance scales are not bad. They are misused. The device is a rough estimator with a well-characterized 3 to 4 percentage-point error band that swings almost entirely with hydration and other confounders you can control by disciplined measurement. Under a strict protocol — same time, same conditions, weekly averaging, quarterly DEXA calibration — it becomes a serviceable trend tool. Under an unstructured protocol it is a mood-swing generator.
If body-composition change is the actual goal, the highest-leverage steps are the boring ones: a bathroom scale you trust for weight, a fabric tape measure for waist circumference at the navel, a mirror for genuine visual comparison, and a DEXA scan twice a year for calibration. A BIA scale added to that stack, used well, sharpens the picture. A BIA scale used casually blurs it.
When we add one to the catalog and score it, we will publish the rubric breakdown transparently, the way we did for Dexcom Stelo and Withings ScanWatch. Until then, know what the tool is measuring, and set your expectations to match the physics.