Smart Scales and Body Fat: What to Trust

Stand on the metal pads barefoot and the scale sends a small alternating current up one leg and down the other — far too weak to feel. Lean tissue is full of water and electrolytes, so it conducts easily; fat resists. The scale measures how much the current is impeded and gets one number: your body's resistance.

Resistance is not a body fat percentage. To bridge the gap, the scale applies a regression equation - a formula built by measuring a group of volunteers with a reference method, then fitting a curve through their impedance, height, weight, age and sex. Your reading is that curve's best guess about someone shaped roughly like you.

Two consequences follow, and they explain almost every complaint people have about these devices:

• The equation is proprietary and brand-specific. Two scales given the identical resistance can return numbers several points apart, because they were fitted to different volunteers.
• Foot-to-foot scales mostly read your legs. The current takes the shortest path, so the trunk — where visceral fat matters most — is under-represented. Handheld grip devices have the mirror-image blind spot.

None of that makes the scale useless. It makes it a device that measures one thing (resistance) very consistently, and reports another thing (fat) approximately.

Because the current is really measuring body water, anything that moves your fluid moves the estimate. Common swings, none of them fat:

• Dehydration reads as fatter. Less water, more resistance, higher percentage. A morning after a hard session or a salty dinner routinely adds a point or two.
• Right after exercise reads as leaner. Warm skin, dilated vessels and blood pooled in the legs drop resistance. It reverses within the hour.
• Food, drink and a full bladder shift it, for the same reason they shift the weight.
• Menstrual cycle fluid retention moves it across the month.
• Dry, callused or cold feet raise resistance - a genuine cause of a strangely high reading with no other explanation.

Swings of 4-8 percentage points between conditions are ordinary, which is more than most people's real annual change. Practically: a single body fat reading tells you very little. The same reading taken the same way every morning for two months tells you the direction, and the direction is the part that was worth knowing.

One safety note: manufacturers generally advise against bioimpedance scales if you have a pacemaker or another implanted electronic device, and the estimates are not validated in pregnancy. Weight-only mode, or a scale without body fat, is the straightforward answer.

Search results promise a body fat percentage from a photo or a phone camera. What those apps do is real, but it is not measurement: they extract your silhouette - shoulder, waist and hip widths, and how your outline curves — and feed it to a model trained on thousands of photos whose true body fat was measured by DEXA. The output is a prediction from your shape, not a reading from your tissue.

How well that works depends almost entirely on the conditions. With a fixed pose, close-fitting clothing, even lighting and the same distance every time, the better apps track change respectably — some are built on published 3D body-scanning research. Photographed casually, in different clothes, in different light, they are less consistent than an inexpensive bioimpedance scale, and neither is in the same conversation as a DEXA scan.

So they are worth what a mirror is worth, with a number attached. Two honest uses: as a visual record — front and side photos in the same spot every month say more about composition change than any home percentage does — and as a rough second opinion when the scale's figure looks implausible. What none of them can do is give you an accurate absolute number from a picture, and any app that presents one to the decimal point is overstating what it knows.

If you want a genuine anchor point, these are the options — the full comparison lives in body fat percentage explained:

The line worth reading twice is the tape measure. It costs nothing, needs no equation, and a waist measurement on its own is the most useful home number in this table — it tracks the visceral fat that actually drives risk, and it doesn't care how hydrated you are. If you take one thing from this page beyond weighing consistently, it is measuring your waist once a month.

A single DEXA scan is also more useful than it sounds: it tells you how far your scale's estimate sits from reality, and you can carry that offset in your head afterwards.

Whatever the scale claims to measure, consistency is what turns readings into a trend. The protocol is short:

• Morning, after the bathroom, before eating or drinking. This is the answer to “when is the most accurate time to weigh” - not because morning weight is your true weight, but because it is the moment of the day with the least variation in it.
• Minimal clothing, bare feet. Jeans and a sweater are a kilogram of pure error, and bioimpedance needs skin contact anyway.
• Hard, flat floor. Carpet can shift a reading by a kilogram or more. Mark a spot and always use it.
• One scale, forever. Two scales disagree on weight and disagree far more on body fat. Switching mid-journey puts a step change in your data that looks like a real event.
• Record it immediately. Numbers remembered until the evening drift toward what you hoped.

Then expect movement anyway: a normal day-to-day swing is 1-2 kg, and almost none of it is fat. How often to step on at all is its own trade-off — see how often you should weigh yourself.

A workable home setup uses each number for what it is actually good for:

1. Weight — track the trend, not the day. Averaged across a week, it is accurate, sensitive and honest.
2. Waist — measure monthly. Cheap, stable, and the best single signal of the fat that matters.
3. Body fat — watch the direction over months, on one device, under identical conditions. Never compare it to a friend's scale, a gym scale or last year's different model.

The pairing does the real work. Weight steady while the waist shrinks is recomposition — you would have called that a stalled month if you only owned a scale. Weight falling while the waist holds is worth a closer look at whether you are losing muscle. Neither picture is visible from one number.

A notebook records all three perfectly well, and if that is the habit you will keep, keep it. What paper is bad at is the arithmetic: nobody averages seven numbers by hand every week for a year, and a column of digits hides the trend a line shows instantly.

Weight Journal is free and does that part: weight, BMI and body fat logged together in kg or lbs, a trend line through the noise, charts over any period, goal tracking, reminders, Apple Health sync — so a smart scale that writes to Health can fill it in for you — photos on entries, export to PDF, XLSX, CSV or JSON, and iCloud sync with no account and no email.