When you spend eight straight years designing, developing, and supporting a product by yourself, you gain perspective. You truly feel the consequences of your decisions as they ripple through the years. When you work on a team and are able to specialize, you likely feel some pain from others’ decisions, but you don’t have the full context. And, most importantly, you likely don’t feel the full implications of your own decisions.
When you’re responsible for everything, you see a bigger picture. You hear directly from users about design decisions you thought were smart. You hear from yourself when a dependency upgrade is difficult because of a web of intertwined dependencies. It’s humbling. But you learn.
Flitting from side project to side project is fun and easy, but you don’t get to absorb the full arc of software engineering. And if something isn’t supporting paying customers, you never really feel how it all fits together.
When I started Sifter, I may have had a degree in Computer Science, but I didn’t know the first thing about software engineering. I didn’t know how to deploy. I didn’t appreciate the value of continuous integration or staging environments. Heck, I barely even wrote tests.
But over the years, when releases were delayed because of a clunky deploy process or a failure to write tests led directly to an increase in support requests, I had to change. The feedback loop becomes more direct and the consequences more immediate. With Sifter, those bad habits all evolved and improved with time as I learned, but it took some painful lessons before I really bought in.
So, with hindsight, I’m doing a lot of things differently these days. And, from casual conversations with others, I’m not alone. So what’s changed?
In short. I’ve stopped swimming upstream. And the results? I’m enjoying my work by orders of magnitude more. I’m spending less time wrestling with quirky customizations. I’m iterating quickly and more reliably. Simply put, it’s fun again.
If you ever find yourself in a place where you feel like you’re spending more time fighting with your software than building and shipping it, stop and evaluate which pieces are causing pain and why. Experiment with doing less and using fewer dependencies. I think you’ll be pleasantly surprised.