High variance body composition readings

Let me start by saying that I understand that your body is different from day to day and that you should focus on long term trends/averages rather than single day readings/measurements.

I’ve been using a Withings Body+ scale for a little over a year now and I’ve never been completely happy with it. I just feel like the readings are unrealistically variable. I get on the scale very consistently after I’ve been to the bathroom in the morning and before any intake, but the numbers will vary wildly from day to day (and my workouts from the previous day to not seem to affect anything). For example, from yesterday to today I apparently gained 2.3 pounds, including 1.8 pounds of body fat. Then after measuring this morning, I went on a short easy run (low intensity in cool weather, minimal sweat), peed again, and jumped on the scale again. I had lost 1.7 pounds overall, but 3.5 pounds of fat (woohoo, let’s keep that measurement! :)).

Like I said, I understand that it’s the overall trends that are important. But do others see this kind of daily variance in their measurements? I’m not trying to be a snowflake here, but maybe the withings algorithm doesn’t work on my body? Or the unit is just faulty? Or maybe this is expected?


1 Like

I’ve found the biggest variance is due to fluctuations in glycogen levels. Glycogen will hold onto a water molecule which is heavy and is typically the weight you hear associated with water weight. When I workout, my glycogen levels are at a baseline level. When I notice this weight gain is after a day where I don’t workout. The body prepares for the workout it thinks is coming, but you don’t workout and thus don’t burn the stored energy and release the water. That results in a 3, 4, 5 lb swing, which usually always happens on Tuesday, which is after my Monday day off. Your body composition will be affected on the scale because this is typically seen by the scale as muscle weight. When your muscle weight is high due to this glycogen, your body fat will seem lower, but then when you burn it off, the body fat percentage will go up.

This is also a reason why folks believe Keto and other low-carb diets work is because they exhaust their glycogen really fast, which results in a weight loss initially, and then they gain it back when they start eating carbs again.


Yes precisely! That is exactly my understanding/expectation as well, however (as in the example above) whenever I have an ‘abnormally’ large increase in daily weight, my BF% ALWAYS increases (percentage, not just total mass). Yesterday my workout was a pretty hard weight session in the morning and an easy aerobic ride later in the day (pettit). Even if I did gain 2.3 pounds of mostly water weight (a little high, but not out of the realm of possibility) I would not expect that to manifest as 80% fat.

This is the main thing that puzzles me and makes me thing something is amiss either with my specific unit or with the withings algorithm.

My experience is it fluctuates. I’m not sure if it’s the scale or the body beyond the reasoning above, but I think just due to the nature of how the technology works, it’s inherently inaccurate. You just need to focus on the long goal and not worry so much about the day to day readings.

I have a Withings too and another well regarded one before that. The electric impedance method for estimating body fat / body composition just doesn’t appear to be accurate. If ones body weight is fluctuating multiple lbs between days its clear that you aren’t loosing 2lbs of fat one day (-7k calories) and gaining 2-3 the next day (7-10k calories required.).

If my weight fluctuates by 1+ lbs day to day, its most likely water but the withings thinks it fat. I just ignore it now. And athlete mode says I’m like 10% which is silly.

1 Like

just focus on that and read what @russell.r.sage has written (weight fluctuates!).

Since January 1st I’ve been stepping on the scale first thing after waking up and peeing. My weight varies 2lbs but has been trending down. Its a Fitbit scale and has % fat but I ignore it, however it does vary from day to day, a big swing is x.6 to x.9%.


I used athlete mode on an old Tanita for years and I was rarely above 5%. Good for the ego! It’s just weird that withings rates excess water as fat, seems like a pretty basic thing they should understand.

Interesting. I’ll see swings far in excess of that. Thanks for the data.

1 Like

FWIW we’ve had the Fitbit scale for 9 years and we treat the % fat as a “for entertainment purposes only” feature. It seems roughly in the ballpark and we have no idea of it’s accuracy. The scale itself works well and is accurate if I weigh myself before a doctors office visit and compare to the mechanical scale.

Yeah, that’s what I’m coming around to as well. You’d think that in 2021 we could do a little better than “hey, let’s include this number for shits and giggles, but it’s meaningless.”

It is really meaningless. If you are trying to lose weight, then body composition improvements will follow. And none of us are body builders looking to get to 5-6%, so I’m not really sure why I care what the body fat% is. Do the right things and the rest will follow.


Same to me my Weight can fluctuate about 1kg a day The FetFat and Muscle mass is worse than guessing… I don’t call it a measurement… Some times overnight I gain 1kg of Muscle and the next day it is fat. Also Body+

^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^

Things like “InBody” and all the at-home body composition analyzers use bioelectric impedance analysis. One major flaw is its sensitivity to hydration status and changes. Really sensitive!

Changing hydration status in a single user from very dehydrated to very hydrated could change their reported BF% from like 20% to 14%, no joke. BIA measures how much electricity is “lost” to your fat tissue which doesn’t conduct electricity very well. The more fat you have, the more electricity is lost between their sensors. The more water you have, the leaner it will report you. Likewise, the less water you’ve got onboard, the more fat it will report you as a percentage.

So, with any BIA measurement, if it is corroborated through visual inspection (looking at progress pictures or in mirror, or by what friends/family have been telling you, unprompted), by how clothing is fitting/feeling, or by what we’re seeing on the scale, then we can use it as another piece of semi-reliable data. But if it seems at odds with what we’re genuinely seeing in any of those measures, or it says something like you’ve lost muscle mass over a time period where you were gaining substantial strength in the weight room…then we should probably assume something is amiss.

1 Like

I use the Withings scale and I know as soon as the BF% shows higher, the water % will show lower and vice versa. Every. Single. Time.

1 Like