Fitness wearables: Enough raw numbers; give us smart advice!

I’ve been following fitness trackers and wearables with keen interest for years.  So far, I’ve tried out several different devices from Fitbit, the Withings Pulse, and the Basis B1 watch.  They’ve all been intriguing… but so very, very disappointing.

From my experience, all the current fitness wearables fail to help us identify correlations that could efficiently and dramatically improve our lives.

Sure, a bit of quantification plus fun gamification encourages many to walk more steps or perhaps get more sleep.  But wouldn’t it be helpful to know how doing action [x] strongly correlates with improvements in [y]?  

For instance, current wearables can measure things like temperature, heart rate, and sleep quality.  That means that they should be able to measure correlations between exercise, sleep, and stress levels, for starters.  And from these correlations, they could provide useful and actionable advice such as “You tend to sleep much more soundly when your room temperature is between 69 and 72 degrees.  Your room temperature, however, averages 75 degrees; lowering it may result in better sleep.

Particularly after providing us with more actionable advice like this, consumers would probably be willing to help wearables collect additional useful data.  Imagine if we could say “Okay Watch:  I just drank a small coffee” or push a button or two to indicate that we just ate finished eating a heavy dinner.  From this, our wearable could add caffeine or food consumption to the data corpus, enabling it to offer personalized recommendations like, “We notice that you experience an average of 27% more sleep interruptions on days when you consume caffeine after 1pm.  Quit doing that! 🙂

For instance, on the Withings dashboard, I can see a jumbled plotting of heart rate measurements… taken at different times and in different circumstances.  Thus meaning, well, pretty much nothing.  But my wearables do know when I’ve finished exercising and so could offer me encouraging info like: “Since you began actively exercising again about 7 weeks ago, your heart rate recovery has improved dramatically… from 130 minutes to resting rate down to 78 minutes.  Recovery rate is a strong indicator of heart health [learn more], so keep up the intense aerobic exercising!

Additionally, based on aggregate user data, wearables could provide useful encouragement based on others’ experiences.

For instance, “Other [wearable] users who added two additional intense 30-minute aerobic sessions per week saw an average of 27% sounder sleep within 3 months.” Of course, the userbase would have to be sufficiently large, and obviously privacy would have to be baked into the core of such a program.

Fitbit One

Wearables need to evolve from mindless and repetitive cheerleaders and observers to smart coaches.

Because right now, alas, undoubtedly too many of us are suffering from “move it!” and “congratulations…” fatigue.  For instance, I’m interrupted by the same Basis ‘achievements’ notices on my phone day after day.  I’m sure I could take a few minutes to alter the notifications, rejigger the “challenges,” and so on, but why should I have to?  If I’ve achieved the “Wear it” goal every day for the last 42 days, why on earth would I want to see this alert for a 43rd time?  Okay, Pulse, so I hit 10,000 steps today?  I know that.  I was walking on a treadmill for a couple hours!  Tell me something I don’t know, please! 🙂

Heck, if you’re going to send alerts to my phone, why not prompt me for useful info?  For instance, obviously with opt-in, “Hi Adam, how do you feel?  [Energetic] [Slightly fatigued] [Dog tired]” and then, after a month, note something like “On those days you took a 15 minute nap around 2pm, you were 70% less likely to report fatigue later on those days.”  Or help me determine what my optimal sleep amount is (“Looks like 7 hours of sleep a night is your ‘sweet spot.’  Your reported fatigue increases steeply when you sleep less than that, but remains pretty unchanged when you sleep *more* than 7 hours.“)

Sure, I could keep a ‘personal log’ spreadsheet (and — I’m such a geek — I actually do!), but computers are much better at finding correlations than we humans are, especially when they have scientists adding in sanity checks to possible ‘correlations’ :).

So in summary, our wearables should…

  • surface useful correlations
  • provide actionable advice,
  • highlight encouraging trends
  • all while leveraging the experience of crowds and the wisdom of scientists.

With the amazing improvements in what we can measure in a compact device and the outstanding increases in phone processing power and user interfaces, I’m confident all of this is absolutely achievable.







What do you think?