The Surprise Outcome of Measuring Our Pull Requests Process

What does measuring pull requests have to do with a boat adrift and me forgetting to call a friend on his birthday? The simple act of tracking metrics led to a team engaged in the process and resulted in an incredible improvement in business operations.

A few weeks ago, I forgot to call a friend of mine for his birthday. We used to be close, but over the years we lost touch. Almost daily calls soon became weekly, then monthly.

When I look back, I wonder how it happened. I remember high school, university, then all of a sudden it seemed the relationship had changed, and here we are today.

As much as it feels like a sudden change, I know it was not. It was a slow, steady degradation that flew comfortably under the radar for years.

It was a drift.

You may encounter several drifts in your life. You slowly gain some weight. You don't read as many books as you used to. You lose touch with your loved ones.

Drifts trick you into thinking that everything is fine because yesterday it seemed only slightly better than today, today is not that bad anyway, and tomorrow — will be better again, right?

Measure things

I know only one way of countering the drift: to measure stuff. Weigh yourself regularly. List the books you read. Keep a journal and write about that beer you had with your best friend. How did it feel? And how do you feel today?

This is true in the workplace as well. Think of the processes in place within your team: are there any blind spots? Is there any piece of workflow you feel is okay but you are not really sure?

You can't reliably improve anything you don't keep track of. And in the reverse case, you can't make sure it won't get worse, either.

Last year, in our development team we started measuring our Pull Requests process. It was a seemingly innocuous part of the process, and we had no big expectations of improvement. We were wrong! Just the mere act of tracking some metrics made people way more engaged in the process, and resulted in an incredible improvement over just a couple of months.

The big takeaway for us has been not only that measuring something is useful, but that it is useful even if you still don't know how to improve, or don't have plans to. It makes the topic visible — it surfaces information that slowly builds up the motivation to improve it in the future.

And it doesn't need to be precise. Let's say you routinely ask yourself (or your team) how do you feel today? And keep track of answers on a scale that goes from amazing, very good, all the way down to very bad. It is very subjective and qualitative, but it is enough to create a feedback loop in which you start drawing correlations between events with metrics, making you naturally gravitate towards behaviour that improves them.

This is because most of the value doesn't come from the measurement itself. It comes from the intent of measuring. It comes from providing attention to something, signalling that you care.

The End Game

In the long run, the feedback loop reverses the drift into its arch enemy: compound growth. The process drives you in the right direction with little effort, because it has become ingrained in your habits.

At the beginning, things progressively get worse because you don’t spend any energy on improving them. By setting the right process and habits, you eventually come to a place where you spend little energy, but things improve on their own.

Previously published at https://refactoring.substack.com/p/the-drift-