How many of these sound familiar?

"What gets measured gets improved."

"If you don't know where you're going, any road'll take you there."

“Every line is the perfect length if you don't measure it.”

"I have been struck again and again by how important measurement is to improving the human condition."

These are all quotes about the importance of metrics in evaluating your next move, whether in life or in business. These quotes are true, we know it. We resolve to track the number of steps we take in a day, or exactly how many hours we spend working instead of browsing social media. We collect data as if we were doing a scientific research project on ourselves. And in the end, we learn. . . not much.

At ProjectLocker, we have things that are going well, and things that we want to improve. We have been using KISSMetrics for a while alongside some homegrown metrics to understand things like how fast we're growing, where and how we're gaining customers, and our conversion rate. As we met to discuss what needs fixing, we discovered that we weren't finding our metrics very useful.

Tools like KISSMetrics are great at what they do, and they're easy to set up. However, many metrics tracking services encourage you to create lots of different kinds of events. Don't you want to know why people who went to the Subversion Import page didn't do an import? Or what percentage of users clicked the Remote Deploy link, followed by Notifications? The fact that you can create an event for anything can cause one of two problems. Either you are paralyzed by the choices and end up not analyzing anything, or you track everything and then find you can't make sense of the data.

I decided to rename our metrics to help me make sense of them. Instead of things like "Visited Site -> Signed Up" or "Signed Up -> Completed Onboarding", I decided to follow the Jeopardy! format, and answer in the form of a question. "Visited Site -> Signed Up" became "How many people who visited the site signed up for paid accounts?". "Signed Up -> Completed Onboarding" became "How many people who signed up successfully uploaded a project?"

When we changed the metrics, it became obvious what was important to measure. Asking human questions instead of writing abridged notations more suited to a 90s database field gives us a bit of a figurative shake-up every time we look at the data. It becomes very obvious what's most important, as the important questions rise to the top and unimportant or convoluted questions sink to the bottom.

For us, we learned that the answer to conversion rate, "How many people who visited the site signed up for paid accounts?", is where we can get the most leverage. Other teams may need to focus on trial conversion rate, or on churn. Regardless of what your one big thing to tune is, asking yourself simple questions will teach you more than any number of pieces of data taken out of context.

So what are you measuring? And what is the one thing you should be tracking that you're not?