What Building vynl Taught Me

"looks nice, doesn't work"
— A friend reviewing an early version of vynl.

At the time, it stung. It was also the most accurate feedback I got.

vynl started with a simple idea: Spotify Wrapped, but better. It's a web app I built on top of Spotify's data. Instead of doing a flashy “here's your year” thing, it tries to look at your listening habits over time. Stuff like patterns, phases, and how your taste slowly changes without you noticing. It's not a Spotify Wrapped clone, and it's definitely not a startup. It was mostly me overanalyzing music stats instead of doing literally anything productive.

I wanted something that didn't just show you your stats once a year, but actually helped you understand your listening habits over time. Patterns, phases, taste changes. It sounded clean in my head. Building it was not. At all. This isn't really a success story. It's more of a log of things I underestimated while building my first serious project.

I had a clear idea, not a clear execution.

The core idea of vynl never changed, but what changed was how I tried to build it.

The first mistake I made when developing vynl was not having a clear plan. This made it easy to add unnecessary features that I thought were “cool”. At several points, I built systems for a future that didn't exist yet. I optimized for scale without users. I planned for growth before I had stability. Basically, I was preparing for success I had not earned. It felt necessary at the time. It was not. I was setting vynl up for 1000 users when I had barely 10.

Open-sourcing it wasn't the plan

Open-sourcing vynl was never part of the original plan. It was more of a last straw. The restrictions of the Spotify API made it almost impossible for me to publish vynl. So, I open-sourced it. Anyone who wants to use it can self-host it, tweak it, or break it in new and creative ways. Not ideal, but better than letting the project die quietly on my laptop.

What I'd do differently if I started over

If I rebuilt vynl now, with all the lessons I've learnt, I would start with a minimum viable product and focus on user feedback before adding features. I would listen to what users want, not what I think they need. I'd focus on one clear use case. I wouldn't care so much about polishing everything. That would be a problem for future me.

What didn't go wrong

I built something I liked. Yes, the process of building it was, let's just say, not the best, but with every mistake, I learned more. I learned more from building vynl than I ever could have from a course or tutorial. It wasn't the success I hoped it would be.

And honestly, I'm fine with that.

▰▱▰