Here at FastCo.Labs, we've been pondering how to cover wellness—you know, fitness, nutrition, stress management—in a way that makes sense for people on technical teams. The idea got me wondering whether all these "quantified self" apps were actually a useful way for people to get into fitness, or whether they were just toys.
As it turns out, he hadn't looked at the data this way, but his reply was encouraging: "I bet if I went back to the database team and had them run some queries to try to look at kind of trends in the aggregate," he said, "we could figure out whether their usage patterns remain consistent, versus improving either in frequency or in distance or in pace." It's not a controlled experiment, to be sure, but it could be interesting.
If users jumped in fitness and then flatlined, then the app probably wasn't helping their fitness goals. If they continued to improve, then at least it was evidence something was working. Even without a control group of non-app-using runners, we could at least get an early peek at whether apps like Runkeeper are actually conducive to continued improvement.
A few days later, I got an email from Runkeeper Director of Analytics Sandeep Hazarika, with some encouraging results. The performance metrics that he analyzed: average total distance tracked per week, average longest distance tracked per week, and average pace per week.
This analysis looks at January 2013 registered cohorts who had at least one trip in each of the months of February, March, and April (Note that three consecutive-month trippers used to be tracked as "Engaged" users.)
Average Total Distance: The improvement from week 0 to week 11 was 26% (from 3.2 miles to 4.1 miles)
Here's the data in graph form:
Average Longest Distance: The percent improvement from week 0 to week 11 was 20% (from 3.8 miles to 4.6 miles). It is interesting to note that the improvement in longest distance plateaus around week 9, and there is a slight decline after that.
Average Pace: The percent improvement from week 0 to week 11 was 11% (from 12.93 minutes per mile to 11.53 minutes per mile)
Jacobs says that up until this point, he had only looked at anecdotal evidence about which features were motivating people—not the overall success of the app itself, versus training without the app.
We know what kinds of things motivate different people. We know that some people are more motivated by guidance, other people are more motivated by things like social accountability, or some people are motivated by free stuff, so by tying in with rewards systems, we take advantage of that. That's really motivational to them. For example, we know that once you have X number of friends, you are going to be far more engaged than if you didn't.
But back to the original question: Are there commonalities between iterating upon your software project and iterating on your body? Mike Oliver, Runkeeper's lead iOS engineer, thinks there might be—at least as far as milestone thinking goes:
As engineers, we have a very specific mind-set. We're scientists at heart, and we look for specific evidence to prove something and then use that evidence to do something. When I go out and run, I expect to run a very specific set of miles for a very specific goal and purpose. I don't know that's necessarily true of non-engineering mind-sets. Somewhere in the back of my mind, the engineer screams "this number of miles is equivalent to exactly this number of potato chips or cookies you can eat later on!" If you just push past an extra street, you can visualize the cookie right there. I think a lot people visualize like that, but engineers even more so.
Are you an engineer and also a fitness buff? We'd love to hear your thoughts about the topic. Tweet me @chrisdannen and let me know.
[Image: Flickr user Ed Yourdon]