Are smart scales accurate? What UK and hospital tests found, and how to check your own

The short answer: weight versus body composition

A smart scale does two jobs, and it does them with different reliability. It weighs you, which is a direct measurement. It also reports body fat, muscle or water, which are estimates calculated from a weak electrical current. The first is close to the truth. The second can be well off.

Saga’s headline puts it plainly: smart scales are very accurate for weight and significantly less so for body composition.¹ This page sets out the evidence, explains why the second job is harder and gives you a way to test your own scale.

How does a smart scale estimate fat and muscle?

The BHF explains that a smart scale sends tiny electrical impulses through the body and measures how quickly they return. Current passes more easily through water-rich tissue such as muscle and blood than through fat or bone, so the speed of the return gives a clue about what the body is made of.⁶

A European review (ESPEN) says the method can determine fat-free mass and total body water in people without significant fluid or electrolyte abnormalities, when appropriate equations are used. It also says the method suffers from a lack of standardised method and quality control procedures.⁵ So the result depends on the maker’s equation, and the equations are not the same from scale to scale.

For the wider picture of what the figures mean, see what is body composition and body fat percentage.

What did independent tests find?

The Saga and University of Bath test

Saga compared eight commercially available smart scales with a Tanita SC240MA professional body composition scale, which has a recommended retail price of around £1,500. Four men and women in their 40s and 50s took part, and skinfold callipers were a second reference. It was published on 27 May 2023.¹

On weight, every scale was more than 99% accurate on average. All except the Withings Body Scan were within 0.3 kg (11 oz). The largest single gap was 0.6 kg (1 lb 5 oz) on a Withings Body Scan, which Saga suspected was an error, and it was still within 1%.¹

On body fat, the picture changed. Two of the eight, the Garmin Index S2 and the Fitbit Aria Air, gave no body composition figures. The others differed from the callipers by 6.75% on average and from the Tanita by 7.12%. The Withings Body Scan was best at 4.3%. The Eufy Smart Scale P2 Pro and the Xiaomi Mi Body Composition Scale 2 were the furthest out, at 11.74% and 10.68%.¹ Saga also found that the Tanita and the callipers differed from each other by around 2.75%, which tells you that the references themselves do not agree exactly.¹

That one person’s numbers differ by +6.5, +15.7 and +19.8 percentage points from the callipers (17.2 minus 10.7, 26.4 minus 10.7 and 30.5 minus 10.7). The Xiaomi and Eufy readings are 2.5 and 2.9 times the callipers figure, which Saga describes as almost three times for the highest.¹ Saga also noticed that volunteers with lower body fat seemed to have slightly larger discrepancies, and it asks whether scales find low body fat harder to estimate.¹

Saga is careful about its own limits. The sample was small and the readings were not repeated, so the numbers are not definitive.¹ Do not read them as figures for every scale or every person.

The French hospital study

Frija-Masson and colleagues compared three smart scales with DEXA in 53, 52 and 48 hospital patients.² Median absolute errors for body weight were 0.3 kg, 0 kg and 0.25 kg. For fat mass they were 2.2 kg, 4.4 kg and 3.7 kg. The authors concluded that smart scales are not accurate for body composition and should not replace DEXA in patient care.² They also found that fat mass error was associated with weight on two of the scales, so errors were not random across body types.²

A professional device, for comparison

A study of a professional multi-frequency bio-impedance device, not a home scale, compared it with DXA in 102 adults. It overestimated fat-free mass by 2.33 ± 2.80 kg (3.6%), and hydration within the range studied did not significantly change the differences.⁴ This shows that even professional impedance equipment carries a margin of error.

Why does your reading change between morning and evening?

Because impedance depends on how much water is in your body and where it is. The BHF warns that smart scales can overestimate body fat if you are dehydrated, and advises weighing at the same time of day, under the same conditions and on the same scales, for example in the morning after emptying your bladder.⁶⁷

Chutex says on the Vita page that its bio-impedance estimates are indicative and vary with hydration, meals, recent exercise and time of day, and it advises avoiding weighing immediately after exercise or a large meal. The page also says Vita does not claim clinical-grade accuracy like DEXA.¹³

How large is each effect? No source read for this guide gives a figure in percentage points for a home scale. The professional-device study found no significant effect of hydration within the range it studied,⁴ so the honest position is that the direction is known and the size is not.

Where the scale stands matters too. Which? found that soft flooring will throw off the readings and advises a flat, hard surface.³

How to check your own scale in five steps

A home test can check the weight function. It cannot check body fat, because you have no reference for that. Which? suggests checking against a known weight, such as a bag of sugar or a dumbbell.³ Here is a way to do it.

1. Put the scale on a flat, hard floor, not carpet.
2. Take three bags of sugar, each labelled 1 kg, and put them in one bag so you weigh them together. The expected weight is 3.00 kg.
3. Weigh the bag three times, stepping the scale off between readings. Write the three numbers down.
4. Work out the gap. If the scale says 3.06 kg, the gap is 0.06 kg, which is 0.06 divided by 3.00, or 2.0%.
5. Repeat on another morning at the same time. If the gap is consistent, you know the scale’s bias. If it jumps around, the scale is unreliable.

What a percentage gap means in stones and pounds

The numbers here are invented to show the arithmetic. A 3% error on a 70 kg person (11 st 0 lb) is 70 x 0.03 = 2.1 kg, which is 4.6 lb, or about 4 lb 10 oz. On a person of 14 st 0 lb (196 lb, 88.9 kg), 3% is 5.9 lb, about 5 lb 14 oz, or 2.7 kg.

Compare that with what Saga found. A scale that is 99% accurate on 70 kg has a margin of 0.7 kg, yet Saga’s gaps were 0.3 kg at most for seven of the eight scales.¹ The weight function is dependable, which is why the sugar-bag check should usually pass.

Makers also state weight accuracy. Withings says Body Comp keeps weight accurate to within 50 g, eufy says the Smart Scale C20 gives precise weight measurements up to 0.1 lb (0.05 kg), and the Vita page states ± 0.05 kg up to 180 kg.¹¹¹²¹³ These are the makers’ claims about weight, not about body fat.

How should you use the numbers?

Saga is reasonably confident that trends over time reflect real change in body composition, for example a body fat percentage falling over three months. It recommends treating smart scales as a guide, and speaking to a doctor about health concerns instead of relying on the scale.¹

• Weigh at the same time, in the same state, on the same scale.⁶
• For tracking weight, the NHS advises weighing once a week rather than daily, because weight fluctuates from day to day.⁹
• Treat body fat as a trend, not a daily verdict.
• Do not use bioimpedance in place of BMI: NICE says not to use it as a substitute for BMI as a measure of general adiposity in adults.⁸
• Pair the scale with a tape measure. See waist measurement and BMI and healthy weight.

Do more expensive scales do better? Saga says the pricier scales tended to be more accurate on body fat, but that even these lack precision, and again that its sample was small.¹ To compare models, see smart scales compared. For deep abdominal fat in particular, see visceral fat.

Who should not use a body composition scale?

The BHF says people with a pacemaker should avoid electronic body fat scales completely.⁷ Makers give their own warnings.

• Garmin says people with pacemakers or other internal electronic devices should not use the Index S2.¹⁴
• Renpho says children under 8 and people who are pregnant or have osteoporosis, a pacemaker, artificial joints, a portable ECG or other metal implants should use its scale for weighing only.¹⁵
• eufy says that with a pregnancy or a pacemaker, the Smart Scale P3 can be switched to a simple weighing mode.¹⁶
• Chutex says on the Vita page that bio-impedance is contraindicated for people with a pacemaker or cardiac stimulator, and that pregnant users can use weight-only mode.¹³

These are statements from makers and the BHF, and this guide gives no medical advice beyond them. If you have an implanted device or are unsure, ask your GP or the team that looks after you before standing on a body composition scale.

Is anyone checking the claims?

The MHRA guidance read for this guide explains that medical device rules aim to ensure devices are acceptably safe and suitable for their intended purpose, but the page does not mention consumer smart scales.¹⁰ No source read says that the NHS or the MHRA validates or classifies consumer smart scales. That is why independent tests, such as those above, matter.

Frequently asked questions