why fitbit is still terrible

Not too long ago, I started using the Fitbit One. I’d been toying with the idea of buying one for a few months, and then a friend of mine purchased one. He told me all about the neat little features (minus the need to constantly sport a bracelet), and not long after, I too was basking in the glory of having my each and every step counted. Not only that, but I got to see how I measured up to my fellow Fitbitters, nearly step by step. I even started tracking my weight, diet and sleep habits, just to see how my diet, steps and sleep impacted my weight. In fact, I am wearing it right now. It’s beautiful having all of this data tracked.

But what is it really doing for me?

Some call this sort of digital fitness tracker a “glorified pedometer,” and in many ways, they are right. Aside from its ability to track my sleeping patterns (when expressly commanded to do so1), I could technically just wear an old-fashioned pedometer, then track my steps at the end of each day in an Excel spreadsheet. I could just use an iPhone app to track my sleeping patterns. I could even share this information with my friends, as annoying as that would be. Yet, the difference is that I wouldn’t. The convenience of this little device means that data gets tracked and I barely have to lift a finger. However, now that I have that data, I am left wanting more.

No, Fitbit, I am not blaming you; it’s not your fault. Quite the contrary — it’s my fault. As a consumer, I have essentially stated, “I need a sleek, out-of-the-way device that tracks my every move,” and then stopped there. What I should be saying is, “I need a convenient device that tracks me well and provides meaningful suggestions for me.”

Yes, just as I could wear an old-fashioned pedometer, I very well could just read the data that my wonderful device gives me and correlate it myself in order to extract some sort of useful information. Again, the difference is that I won’t. Why would I do something that could be done for me? The data is there, after all. Why not put it to work? For example, why can’t it tell me whether I sleep enough to be considered healthy, or when I should go to sleep based on my previous sleeping patterns if I want to get enough rest, or whether my step pattern indicates I’m not getting up from my seated position enough times per day. That’s useful information that could help me do my job better, develop fewer health issues and generally feel less stressed. I’m not asking for miracles here; I’m asking for information in the form of suggestions.

To me, there’s a difference between data and information. Data is unstructured facts and numbers; in essence, simply a log. On the other hand, information tells me what the data means, hopefully in the context of something I can understand. This is an extremely important distinction.

Looking at the Fitbit website, there are a number of items (data) I am currently able to log: Food, Activities, Weight, Sleep, Journal, Heart, BP (Blood Pressure) and Glucose. Out of those eight, one is technically automatic, and the other (sleep) is nearly automatic. The rest require external input. I’m not saying they should have packed all these features in as automatic; I understand the importance of doing a few things really well as opposed to doing many things not so well. However, I also understand incremental (read: slow and steady) changes for the benefit of a product.

Take the “Journal” feature, for example. On the Fitbit website, this screen asks me what my mood is, whether I’m suffering from allergies and asks me to write a nice little entry. In the blog entry they wrote on the subject, they highlight its usefulness:

Now you can track your mood, allergy level and generally record your daily observations with the new fitbit journal feature. And use the information you record to identify trends and correlate your sleep and mood with other factors that may be affecting your activity level.

Notice they don’t mention anything about their service correlating sleep, mood and other factors. They’re talking about me doing the work. Me. The guy who won’t even bother tracking my steps in an Excel spreadsheet. Again, I am understanding here. After all, how are they to get that sort of data out of the existing device? Unless it has a heart rate monitor built in that I don’t know about, I’d imagine the task next to impossible. Yet, I have seen small devices that track heart rate, so why am I hearing nothing about plans for this feature?

I’m not saying you can directly derive mood from heart rate2, but since the Fitbit is already talking to my iPhone, why not detect sharp changes in heart rate and then prompt me for my current mood? And sure, I can write a journal entry later if I feel like it. What’s important is that, at my discretion, mood can now be linked to heart rate, along with my sleeping and step patterns. Now, on top of the other examples given earlier, I can receive even better information: how my sleep patterns correlate to instances of anxiety or stress in a given day, how sleep length impacts happiness, how my level of activity impacts happiness, etc. I could go on with these examples. Why leave the correlating up to me? Sure, give us access to the data; I’m sure some people could come up with charts and graphs of their own to provide themselves with meaningful information, but not me.

And, of course, I know Fitbit works with the Aria Scale to track weight. I haven’t used it yet, so I don’t know whether it yields the kind of information I’m after, but given the rest of the Fitbit experience, I’d bet not.

If I’m not limiting myself to the real world, there is a magical device that tracks me, runs algorithms to determine optimal progress, then interacts with me non-judgmentally to work toward goals. If that means automatically tracking my daily activities, moods, heart rate, blood pressure, sleep patterns, glucose and cortisol levels3, so be it. That’s the future I want to live in, even if it takes a few device iterations to get there.

On the other hand, Fitbit, what’s stopping you from crunching some numbers in the mean time? While I’m waiting for the next tracking feature on your amazing little device, why not work on the information you’re providing to us through your service? I mean, the data is there for the taking. In my case, you’ve got three months work with.

But maybe I’m oversimplifying things. To me, it seems like you could do what Google does with search data to my fitness data. Why not run my seemingly simple information through a number of different algorithms to give me meaningful, tailored suggestions? It seems like the next logical step. And while I am still loyal to Fitbit, someone is bound to come along and do this even if you don’t. In fact, there’s really nothing stopping another company from using the data provided by Fitbit to grant my digital wishes.

Now there’s an idea.

 


 

  1. In order to put my Fitbit One in Sleep Tracker Mode, I must wear a special wrist band and, when I am about to sleep, press its little button for ~2 seconds to signal sleep readiness. Once I wake up, I hold the same button for ~2 seconds to signal the fact that I am now awake.
  2. In a study titled The Relationship Between Heart Rate and Mood in Real Life conducted by Derek W. Johnston and Pavlos Anastasiades, it was stated that, “Heart rate related to emotional state in very few subjects when time-series statistical methods, which take into account the autocorrelated nature of the data, were used.” To me, that says you can’t consistently determine mood from heart rate. Just saying.
  3. Have you heard of Cue? “Cue is a revolutionary new device that connects you to your health at the molecular level. Simply load a cartridge and add a sample to access deep information about your body, on demand and on your schedule. Discover every day how activity, food, and sleep shape your body’s story. And achieve meaningful, daily improvments to tell a new one.” Now that’s what I’m talking about.

6 thoughts on “why fitbit is still terrible

  1. Why leave the correlating up to you? Because you are the only person who understands your data well enough to make meaningful correlations.

    Fitbit’s service is designed as a motivational tool. If you want to get answers to specific questions (e.g. how does exercise affect your sleep?), you’ll need to take your Fitbit data elsewhere.

    There are some apps that use voice samples to detect your stress level (look for StressSense or AIRS); haven’t tested this myself yet.

    1. Eric, thanks for the response. I just want to start my reply by clarifying my title — it was meant to grab attention. I actually love my Fitbit.

      Why leave the correlating up to you? Because you are the only person who understands your data well enough to make meaningful correlations.

      I’m sure there are things specific to me that a service like Fitbit wouldn’t be able to provide, but that doesn’t mean they couldn’t give me meaningful correlations between standard sets of data.

      Fitbit’s service is designed as a motivational tool. If you want to get answers to specific questions (e.g. how does exercise affect your sleep?), you’ll need to take your Fitbit data elsewhere.

      I agree that it’s designed as a motivational tool, and part of my post here was specifically about exporting that data to a service that does the correlating elsewhere; however, from a consumer perspective (mine specifically), it’s the next step. If we had a sort of evolution chart from a simple pedometer to what might be the next adaptation, from my perspective, it would be something with simple functionality, yet more feedback.

      Seeing the sentence, “you’ll need to take your Fitbit data elsewhere” feels like a nail in the Fitbit coffin. Is that really all they ever plan to offer?

      There are some apps that use voice samples to detect your stress level (look for StressSense or AIRS); haven’t tested this myself yet.

      Right! I’ve also used the Cardiio app, which can (apparently) detect heart rate from just by reading your face, or using your phone’s camera and camera flash.

      My objective with this post wasn’t to rip into Fitbit; on the contrary, Fitbit is generally known as being awesome (at least in my circles). The point was to say, yes — you’re awesome, but since I’m one of your loyal supporters, there’s still room for improvement.

      I checked out Zenobase — nice app, by the way — and I can see where you’re coming from, but I don’t necessarily agree 100%. I think Fitbit can (optionally) provide some of the type of information I mentioned in the post and those things would apply to the general population.

      For other data, there’s always a service like yours, or just old-fashioned correlating via Excel.

      1. I’m sure Fitbit would love to be able to provide you with actionable insights; the problem is that the process of coming up with useful insights can’t be automated.

        1. Eric, I disagree. I think useful insights could be automated for a small common set of data, assuming all the requisite data is tracked. Presenting that information in a tailored way could pose a challenge, I suppose, but if there are businesses based on it, I don’t think it’s impossible.

          Take addapp.io for instance; they allow you to correlate data from a range of sources, then perform “experiments” to test whether there is a positive correlation between data. So far, I agree with you. The next step is that other people can view the same experiments and if they have similar data, utilize the same experiments on themselves. So in my opinion, that works. You could do it this way and then also provide automated correlations for qualifying data. People could shop for experiments the same way you shop for plugins on WordPress. This is the kind of thing I’m talking about.

          1. Data analysis is an interactive process; naively correlating unfiltered data from multiple sources rarely produces useful insights.

            I agree that sharing the setup and results of experiments is valuable. That’s the bread and butter of QS meetups, and one of the goals of personalexperiments.org.

          2. Eric, thanks for your feedback. I am still not 100% convinced that certain correlations could be made for people tracking a common set of statistics, but I have nothing substantial to back up my doubt, so I’ll continue filling in my own knowledge gaps for a potential future apology post for Fitbit. :)

            My hope for Fitbit and similar health-related devices are that they continue to improve and offer us better feedback (even if that means as a user I have to be more manually involved). Again, I’m a big fan of what Fitbit has done, and I hope it’s just the beginning. Eventually, I’d like to see devices that work with me to achieve better health. Not just me, but also those that look after my health (doctors, dentists, etc.).

            Privacy concerns aside, I’m sure I’m not alone here. Of course, everything of this nature has to be done with security in mind, but I think we have a lot to gain from constant, yet non-invasive and non-judgemental, intelligent feedback about our life choices.

            Forgive my ignorance about QS, but from a quick Google search, it seems like you’re talking about a thing called ‘Quantified Self.’ Before this moment, I hadn’t heard of this at all! This seems very interesting indeed, and I’ll be spending some time looking into it.

            Thanks again.

Comments are closed.