Solving Your Product’s Retention Issues — How we did this in the early days of RunKeeper

Distribution and retention are two separate problems startups need to solve. Scaling your distribution channels requires you to know why people are coming to your product. Growing your active user base requires you to know why users are coming back for more. But do not confuse the two. Rarely do users stick around for the same reason they came in the first place.

Finding product distribution is always your number one problem but the retention problem can be even more challenging as you can’t simply use a canned Google analytics report to answer why users are sticking around.

Facebook & Twitter
During trips to Silicon Valley, I got to learn from Facebook and Twitter veterans who shared their stories on how they solved their respective retention problems. In the early days of Facebook, users signed up after receiving email after email from their friends who were already on the site. But, according to the early Facebook employees I’ve spoke with, they discovered that users would stick around after seeing photos of their friends ..actually, off the record I was told it was photos of their attractive friends (although I don’t know how they figured that out). Similarly at Twitter, users came to see why everyone was talking about this new thing called “Twitter”, but the average user would only return once they followed a celebrity who filled their feed with insight into the lives of the famous.

RunKeeper

Adding to the lessons learned from Facebook and Twitter, I think it might also be helpful to know how we went about solving the same retention problem in the early days of RunKeeper. At RunKeeper, we knew people were signing up because our app was a low cost (free) alternative to expensive GPS watches ($0 < $200). Users saw their friends posting individual running activities and a map of the route to Facebook and decided to download the app to give it a try. We knew this very early on, but we didn’t know, for quite some time, why they stayed.

Finding out why they stayed was a long process that took years to figure out. You see, we never successfully built a large analytics platform at RunKeeper for the first 4 years I was there. We could track our incoming channels thanks to a tool built by an intern one summer, but it wasn’t until I left the company did they ever have the ability to truly dive into the analytics behind user behavior. By the way, I’m totally jealous I never got to see this data.

Instead, we built new features based on listening to user feedback and watching which features seemed to be popular on competing products. By doing so, we built a race platform, a training program platform, and a suite of social networking features. We then listened to user feedback and watched our database to see which of these features were being used the most. The race platform was hugely successful but people only used it to declare which races they were training for. Other features of the race platform, such as spectators watching the races live, were barely touched. The training programs (then called “FitnessClasses” at the time) were not very popular except for the losing weight program.

From this, it became our suspicion that people were continuing to use RunKeeper to achieve a goal such as running a race or to lose weight. They came to RunKeeper to track an individual fitness activity, but may have been sticking around because we were helping them progress towards their personal fitness goal. Seems obvious now, but not obvious at the time.

Taking Twitter’s lead of placing the “pick a celebrity to follow” walk-though at time of sign ups, we decided to investigate placing a “pick your goal” walk-though right after sign up as well. This feature was launched after I left RunKeeper but it still seems to be a main focus of the product today.

Retention is a hard nut to crack. It takes time, patience, and the ability to really listen to why your active users are using the product. You can only try to solve this problem once you have started to solve your distribution problem, but once you do, you’ll really have a great product that will be used and loved by a large active user base.