Sleep data - Comparing Hello Sense, Fitbit Charge and Bellabeat Leaf

Thought it might be fun to run a comparison of sleep logging on a few of my devices.  

For reference, I wore the Fitbit Charge on my left (non-dominant) wrist, the Bellabeat Leaf clipped to the neckline of my night vest and the Hello Sense is of course clipped to my pillowcase. I sleep primarily on my side/smushed partially face-down and I'm quite a light-sleeper.

In terms of data points I've recorded, here's a break down:

  • Sleep score - This is a number the software gives, generally a percentage of total sleep time.
  • Total sleeping time - The time the software said I was asleep in some way (with caveats, see next item).
  • Deep sleep - This seems to vary by system, some counting what I presume they're inferring to be REM time, but it looks like Fitbit is counting this as all the time I was "asleep" with "total sleeping time" being time in bed total, asleep or not.
  • Times woken - Exactly what it says.  I wake at least once a night, every night, for sure.
  • Fell asleep - The time the software believes I, well, fell asleep.
  • Wake up - The time the software believes I was first awake in the morning.
  • Out of bed - The time the software believes I physically got out of bed.
  • A ? indicates - The software doesn't offer a value that fits the category.

So, let's have a look at the numbers!  

I've got 3 full days of data, and one partial. Why so little? I forgot to record the Sense data each morning for a few days and it turns out you can't go back and look at previous nights.  Major fail for the Hello Sense right there.  I'm also missing a night of the Leaf's data because I forgot to put it on. Update: The friendly Hello Sense people just emailed me to let me know I can go back and see previous nights data! The gesture was a little non-obvious - but a slide right will go backwards. 

 Fitbit ChargeHello SenseBellabeat Leaf
MondaySleep score96%77-
 Total sleeping time7h54m6.7h-
 Deep sleep7h2m2.5h-
 Times woken20-
 Fell asleep23:560.31-
 Wake up07:5607:12-
 Out of bed?08:35-
TuesdaySleep score93%8197%
 Total sleeping time10h11m8.9h9h24m
 Deep sleep8h49m2.7h?
 Times woken201
 Fell asleep22:4123.5823:07
 Wake up08:518.5208:42
 Out of bed?8.48?
WednesdaySleep score97%7898%
 Total sleeping time9h36m9.9h8h58m
 Deep sleep8h56m4.8h?
 Times woken302
 Fell asleep23:3923:1700:19
 Wake up09:1409:1309:39
 Out of bed?09:24?
ThursdaySleep score95%8176%
 Total sleeping time10h7m7.6h5h32m
 Deep sleep9h35m3.1h?
 Times woken200
 Fell asleep22:4323:0601:36
 Wake up08:4906:4207:08
 Out of bed?09:01?


On Wednesday, the Sense alarm failed and didn't wake me up, so I overslept and on Thursday only 1 of two Sense alarms went off (I reported the outage, and I think they know what it was and fixed it). Thursday night I *definitely* got up at least twice.  I had a Leaf alarm set every morning, and although I felt it go off once (because I was already awake), it was never the thing to wake me up. The Fitbit alarm works well and has yet to not wake me up when used.  For reference, my first alarm in the morning goes off at 8:00am (lucky me) and then a second to get me up at 8:25 if I haven't shifted by then.

Which feels more accurate?  

Based on what I know about my own sleep habits and routine (don't judge my lazy ass), the Fitbit Charge data fits best and I'd take that one as feeling most accurate - not having any proper professional equipment strapped to me, I can't be certain, of course.  

Comparing the BellaBeat Leaf to the Fitbit data, it does seem to be pretty far off the mark - often not noticing that I've fell asleep until after the other systems.  It was also the weakest in terms of the amount of data and breakdowns in the 3 systems. 

The Sense seems intermittently accurate and is prone to believe I'm awake a lot earlier than I generally am.  I'm very surprised that the Sense hasn't noticed when I've been out of bed at night, but I wonder how much bed sharing effects that (it often just notes "you and your partner were both restless" at certain times that I suspect may be when I've been up or awake).  

Having said that, I don't think any of them are accurately recording how much time I spend in quality sleep, but I don't particularly expect a system that relies solely on tracking movement to be able to do that.  Make of this data what you will , but the bottom line on sleep data by fitness trackers should probably be to take what they tell you with a large pinch of salt.