Gadget Review: Fitbit "Flex"

[hrj]

It's probably premature to do a review of a product I only started using today, but this will probably be added to.

I'd been looking vaguely at various "fitness tracker" devices that sync to your smartphone but I hadn't yet found a sales person who actually knew the mechanics of how the things worked and I was having a hard time translating the somewhat grandiose claims ("tracks how many calories you burn", "analyzes your sleep patterns") with the lack of any obvious mechanism for true biometrics (e.g., heart rate, brainwaves). And, in fact, as far as I can tell, the vast majority of entry-level devices rely solely on accelerometers. They track how and how often you wave them around due to their positioning on some part of your anatomy.

So I hadn't yet found the information and/or features I wanted in order to be willing to shell out the non-trivial cost … until my employer decided that this year's inducement to participate in their "healthy living" program was a free Fitbit device. Well, "free" -- that's worth the experiment. So I signed up and selected the "Flex" which is the wristband style. Being the perverse sort of person that I am, my first day's experiment involved putting it on my ankle instead. (Since it's designed for a wrist, it only fits on my right ankle -- the one where the sciatica has caused enough muscle wasting at the ankle to slim it down sufficiently.) My theory is that since -- as I've determined -- the device relies solely on an accelerometer, it may measure rather differently when tracking foot movements than hand movements. And since my major exercises (bicycling and elliptical) involve functionally immobile wrists, I expect that tomorrow's experiment -- with essentially identical activity but the band on the wrist -- will show substantially lower numbers.

The nature of the measurement also ensures that vigorous exercises that don't involve high-impact don't count as "very active" for the devices metrics. (The FAQ outright states, "your active minute count will be lower for activities that are not primarily step-based, such as weight lifting, cycling, and rowing." Hmm, weight lifting (check), cycling (check), and rowing (check). Also, I would add, elliptical (check). No doubt due to the smooth motion. In other words, "If you do an activity that will fuck up your knees, then we'll give you extra points, but this actually has nothing to do with intensity of effort." So my day included at least 40 minutes of vigorous cycling and slightly less than 40 minutes of vigorous elliptical but my Fitbit only credits me with 11 "very active" minutes.

On the other hand, the device does seem to count all my leg movements as actual steps (although it only counts steps by the leg it's attached to -- as I assume it only counts as a "step" the motion of the wrist it's attached to). I know this because the elliptical counts both feet and gives me my full 5000+ steps for my workout but in the same time the Fitbit only credited me with ca. 2500. OK, fine, it's a relative measurement. But it isn't actually counting "steps".

The "sleep tracking" is similarly less useful than the hype. You start tracking your sleep by tapping on the device to say "start sleep tracking" and you stop by tapping again to go out of sleep mode. So let's say I'm having bouts of insomnia and repeated waking. I will be able to track exactly how much sleep I actually get because at that point when I finally manage to drop off … I will wake up sufficiently to tap the device, ensuring that I will then lie awake for at least another half hour. Or am I missing something? (It does track movements during sleep, so I guess tossing and turning will be tracked.)

So let's see what Fitbit claims today's numbers are. I took 11,100 steps, covered 4.63 miles, had 11 "very active" minutes, and burned 1810 calories (they include basal metabolism in this). I'll accept their "step" count with the above caveat, but bicycling and elliptical combined I know totals over 13 miles (so they're calculating how far I would have gotten if I low-impact walked 11,100 steps). And as noted I actually had more like 80 "very active" minutes. Based on calorie guides for specific activities, based on specific effort levels, my high-impact activities estimate out at about 1100 calories (not including low-impact stuff and basal metabolism). And I predict that tomorrow's tracking (on the wrist) will even more drastically under-report my activity.

So unless your primary exercise activity is running, don't count on a Fitbit to track your actual exercise efforts. But if you calibrate it against whatever your target activity routine is, it may be useful for tracking whether you hit that target.

Comments

[kareina.livejournal.com]

For the sleep thing--I don't have anything like a Fitbit, but I do use the MyLogs app to track lots of stuff manually. I tell my phone when I lay down at night, and when I wake up in the morning, and just accept the fact that while I may have been in bed 5 or 8 or whatever hours, I won't have actually slept all of that, but it was, at least, rest when compared to doing physical or intellectually challenging work. Sometimes I tell it I am laying down a second time because we got to talking and stayed awake, others I am too tired to actually push the button, so I look at the clock and try to remember the time I finally settle into a proper sleeping position so I can do the data entry when I wake up.