How accurate are wearable devices for fitness? And does it matter?
There’s a lot of research out there to suggest there are issues with actual accuracy of what’s being recorded, and variances between different devices and what they measure too. If you use different fitness bands simultaneously you will no doubt have noticed some of this.
I’m a runner, and into tracking and measuring stuff, so I decided to test out my own wearables and how they compared in data gathered. And where better place to test the lot than run the San Francisco Marathon (2014) with the lot.
I now normally run using the Strava running app to track my performance (sometimes using the Strava Google Glassware at times) and MapMyRun (I’ve given up on RunKeeper purely to keep the app level down, but I’ve used it heavily too ) on two iPhones or a Samsung III, and I also use a FitBit Flex and Misfit Wearables Shine.
As you can see, there are some variances evident from my trial. Granted in some cases, there is closeness in the data recorded too. And of course, the methodology is a bit wonky, as I wore the FitBit and Shine the whole day whereas I only turned on Strava and MapMyRun for the Marathon, and the metrics ain’t always the same either.
I should say of course, that on top of my official marathon chip time of 3:35:05 for the 26.2 miles, I did put in a 14 mile recovery run that night too and that showed up on the FitBit and Shine figures. So that’s over 40 miles or about 65KM run that day. Yeah, I’m like that…
So, here we go…
MapMyRun (26.86 miles) and FitBit Integration (43.61 miles)
FitBit Flex (70.2 KM or 44.7 miles)
Misfit Shine (62.6 miles total, 3.6 hours running in the marathon)
Draw your own conclusions….
Variances were not as severe as I thought, but there are differences. I have no idea what was going on with the Shine, but whatever.
But does it matter?
Yes, the data matters. But it depends too…
Sure, these kinds of variances are not massive, and the fitness devices and apps devices are consumerized, available widely, and great for tracking progress in a less-than-obsessive way and for motivational purposes. And if they get Joe and Joan Soap out there trying to get a bit healthier and to show off a bit to friends about it, then great. A 10% or 15% loss of data or mismatch with reality isn’t a showstopper here.
But if you’re a serious runner or fitness type, or into the Quantified Self thang, then it really matters. You’ll take this stuff seriously and as a regular user monitoring progress, even small differences can be important. Plus, you’re likely to be comparing performance with others.
Most serious runners I see out there use a Garmin device based on my observations in running a few big races now. Strava and MapMyRun are popular too. Seeing iPhones and Android devices strapped to runners’ arms is now de rigeur, even it’s only for the music, but often there are apps running in the background gathering data as they run. I do see people out there with Jawbone Ups, Nike+ FuelBands, and FitBits and others alongside the serious stuff, but usually that’s just because they’re as easy to leave on as take off.
Critically though, you’re in the enterprise space and using gathered data from users, or employees of your company, for say, negotiating health insurance policies or other perks with suppliers, then it very definitely matters too.
I need to do more research on this, and I will. There are issues to be explored.
Don’t even start me on the inability to exchange data between these devices in the cloud….