My current work project started last summer, as a bit of experimentation. A few of us sat together in a room and started writing down a hypothetical piece of configuration. Within less than a week, we had actually written a prototype-quality piece of software accepting exactly the configuration we had brainstormed.

A few months later however, this project went through a difficult phase where we realized that we actually needed to write down a plan for bringing the software to a stable, usable state. This ended up a painful couple of weeks of planning ahead and realizing that our goal of shipping this was slipping.

I had a conversation about this with my Engineering Director the other day, and I thought what he told me was intriguing. He said that we were struggling with going from what he called “founder mode” to an actual planned development process.

As an analogy, imagine you are painting a wall green. In the beginning, you can just put paint wherever you like. After a while, you can make a break and proudly exclaim that you already painted 30% of that wall. But sooner or later, you will have to check carefully where there is no paint yet. And you will need to go with a smaller brush, so that you can paint the corners properly.

In software development, some people can only do the first phase. They have awesome ideas and they can churn out an early version in record time, maybe even working nights. But they are not good at doing the other 90%: bug fixes, support, identifying areas for improvement, and continual maintenance.

Some other people can only do that second phase. They are very good at maintaining something, sometimes for years, but they might not have the sort of idea to jump-start something completely new. Of course, that’s perfectly fine.

And some people (hopefully) can do both, but they might need to be reminded when it is time to switch to the refinement phase :)

As a leader, if you have people on your team who are the perpetual “founders” (startup, open source, big tech, it doesn’t matter), there is a trick you can pull to keep development going while not antagonizing that “10x engineer”. The trick is to get them interested in some new, shiny problem at the right moment. They might fall in love with solving the next problem and come back wanting to hand off their current thing to someone who will finish building it. And thus, they can do their magic one more time. Even better is when the people who are part of this rochade don’t even notice it happening. –

… Now that I wrote the previous paragraph, I actually wonder if one of my leads has done that with me before, and if so, how many times it has happened!