Find Me


The Fitbit Experiment - 6 Months Later

A few months ago I wrote about buying a Fitbit Ultra for me and my girlfriend. As I descended into the pit of the Quantified Self, I found that simply having an awareness of what I was eating, how much I was exercising, how much water I was drinking and how much sleep I was getting were enough to enact some simple life changes. These changes were small compared to going on a diet or starting up running five miles a day but their effect was dramatic on many levels.

For those of you who don’t know, the Fitbit Ultra is a small device that measures your movement, much like a pedometer but with some additions. It measures changes in elevation, syncs up to the internet and combines smoothly with other devices in the Fitbit ecosystem such as the Aria Wireless Scale and the Fitbit iPhone app.

All of the data collected by the devices, as well as anything you enter into the iPhone app, like food and weight choices, combine to present you with a picture of your day-to-day health.

I started the experiment, after convincing my girlfriend to embark upon this little journey with me, at around 185 lbs. – not overweight by a long stretch for someone of my height. I was in the worst shape of my life, however, and I felt it. I was winded easily and, compared to when I often wandered miles and miles in the streets of New York City, visiting friends, I would feel sore for days after a similar trip. I was also undergoing some big changes at work and stress, combined with lack of sleep, were taking their toll on my body too.

Entering food after each meal into the Fitbit iPhone app, wearing the tracker to bed at night on my wrist (in its handy wristband) and carrying the Fitbit with me wherever I went were the first halting steps towards trying to get some knowledge of what I was really doing on a daily basis.

Some things I quickly discovered:

  • I walk more than I thought I did at work.
  • When I work from home, I get almost no exercise at all.
  • I actually don’t eat too terribly.
  • I drink too much beer, from a caloric perspective.
  • I needed to drink more water. A lot more water.

Some small things I started doing as a result of those findings:

  • Taking the stairs instead of the elevator when I could.
  • Walking to the cafeteria area to get water more often throughout the day.
  • Taking short walks outside at lunch, weather permitting.
  • Parking a bit farther from the front door at work and not minding.

Diet-wise, I do a pretty good job of controlling fat and caloric intake (beer aside). My girlfriend is a vegan so we tend to eat very healthy meals at home. I will order meals that have meat when we got out to restaurants but tend to avoid red meat unless we’re at a very high quality restaurant. Fish and chicken are a general rule although I’m not averse to an occassional Carolina Pulled Pork or Cuban sandwich. All in all, tracking meals with the Fibtit has revealed some of the facts behind what I long-suspected – if I were to reduce cheese and beer in my diet, I’d be a very healthy eater. Unfortunately, those are two of my favorite things.

The Fitbit app has most of the food I eat in its database and, when it doesn’t, I can usually find something similar that reflects a close approximation. Having Fitbit Premium, I can see my average consumption of fat, protein and carbohydrates. That’s a good stat for me to look at given how much cheese I eat and it’s something I’ve never seen so clearly represented.

There are a few failings of the Fitbit ecosystem however. Certainly none that are dealbreakers and I haven’t found a better way to track any of the things I mentioned above, but things that could be improved or managed better.

You can’t see your food intake breakdown in the Fitbit app – only on the website. For someone who is on the go throughout the day and sometimes doesn’t get to their laptop until the evening, or isn’t around the machine that handles syncing with the Fitbit servers throughout the day, that’s a shame. Having the breakdown of fats, carbs and protein available at a glance would not only give some good, actionable information but would also play into the role the Fitbit has served for me during the last few months which is as a guide to subtly affect my behavior. If I knew I hadn’t eaten enough protein on a given day, I could counter it with diet choices later on. Overall, it would end up with a more balanced approach to food due to the immediacy of feedback,

Another failing of the Fitbit system is that it doesn’t seem to be that great for activities other than running or walking. You can push a button before starting an activity and it will track elapsed time but it falls down when trying to tracking more than that. Given that most of my exercise is in the form of walking, I suppose it suffices for now, but when my girlfriend started doing yoga she felt like she wasn’t very well-served by the options presented.

The last thing that needs some improvement is sleep tracking. Sleep tracking is one of my key areas of interest. I was well-known for years among my peers as not getting (and, it can be debated, needing) much sleep. I would sleep four to five hours a night and my body had fallen into a rhythm with that. Over the past few months, I’ve been trying to correct that since many studies have shown that more sleep than 4-6 hours is needed and result in a longer life and healthier body.

The Fitbit allows you to adjust between two levels of sleep sensitivity – low and high sensitivity. The problem is that there is such a wide gulf between the two, it is difficult to even know what they are measuring. For instance, when I had the Fitbit set to high sensitivity, my sleep average per night was 3 hours and 41 minutes. I was in bed at a reasonable hour and rose at a sane time, but my sleep effectiveness was somewhere around 62% due to interruptions and movement throughout the night. After adjusting to low sensitivity, my sleep effectiveness went up to 98% and my average sleep per night was around 6 hours and 32 minutes. Much better, right? But which one was correct? Was it somewhere in between? It is unclear how to make sense of these numbers since there is such a huge disparity.

As a result, I stopped lending any credence to the sleep effectiveness statistic and concentrated fully on “time to bed/rise” data. My girlfriend stopped wearing her Fitbit to bed altogether because she didn’t feel like she was getting any value out of it. Having a way to adjust that sensitivity even further or have a full explanation from the Fitbit folks as to what was being tracked would go a long way to understanding how Fitbit handles the sleep category in general.

Data portability was something I was hoping for but isn’t there yet. I started tracking my allergies on their site every day but realized there is no way to see more than a short window of time from a reporting perspective. As I was hoping to use this data on a yearly trend to track the times I was having worse weeks than others allergy-wise, not being able to extract this data let alone see it made it worthless to me.

The Premium subscription was touted as allowing you to extract your personal data into Excel but the data is aggregated and custom fields weren’t accessible for export. Since that was the main reason to get the Premium subscription for me, I don’t think I’ll continue with the Premium service. If they ever do allow for field-level, time-sliced data points and extraction of custom fields and such, I’ll be the first to sign up again.

Another strange issue is that the synchronization between the Fitbit application on my Mac Mini and the Fitbit website stopped working at some point, presumably around the time I upgraded to Mountain Lion. Restarting the machine seems to fix it but it is a hassle and I wish I knew what was causing it.

Recently, Fitbit announced the release of their newest tracker, the Fitbit One. This device solves my sychronization problem by incorporating Bluetooth 4.0. Using the lower energy consumption of Bluetooth 4.0, the device will sync with newer Macs and iPhones version 4S and up. This might solve the sync issue once and for all.

Six months later, my girlfriend has lost almost 20 pounds and I’ve lost three. I started off well, losing almost 11 pounds but then I started homebrewing. The caloric intake from beer really made the last six months a pitched battle but I’ve still managed to keep a few pounds at bay.

What is interesting is that the weight loss wasn’t a result of some crazy crash diet or drastic life changes. Rather, it was a result of managing food and drink intake and having a better awareness of what you’re doing on a daily basis - a more mindful approach to eating and exercise that affects every phase of your life.

All in all, I’m very pleased with the Fitbit. My girlfriend and I still carry our devices around with us wherever we go every day and knowing, at the very least, our general level of walking exercise is a worthy enough piece of data to continue doing so. Adding more data points, more ways to slice and dice the data and more comprehensive ways to extract atomic data points for my own use would take the entire service’s usefulness to a new level but for now, I’m very pleased with the Fitbit experiment.