A Retrospective on Shutting Down a Side Project After 6 Months
Last summer, I started a side project with a friend and shut it down after 6 months. Here's a breakdown of why.
First, we chose too fancy a stack. k8s, MSA, Kafka... I wonder why we bolted all that onto a service with fewer than 100 users. It was a good way to study infrastructure, but we had no time left to actually build features.
Second, there was no goal. We only had the vague idea of 'launching someday,' but no one decided when or how. Since we both had day jobs, it naturally kept getting pushed back.
Lastly, we only talked about the stack and never about users. If I were to do it again, I'd just write the plan in Notion and build the smallest thing that could ship in a day first.
I spent about 300,000 won, and all that's left is an AWS bill and lessons learned lol.
10 answers
Agreed, if you start by setting up the infrastructure first, 99% of the time you end up not building a single feature lol
If that lesson only cost 300,000 won, isn't that actually cheap? I spent several times that and still haven't sorted it out...
I think the fact that no one talked about users is the real key point. I used to obsess over only building 'technically cool stuff,' too, but I ended up making something nobody used and shut it down. Side projects should start with 'who would use this?' from the beginning, but if the order gets reversed, you just burn time.
Well, I don’t know about that. The problem isn’t the stack—isn’t the real root cause that both got deprioritized under the pretext of having a day job? Even if you hadn’t used k8s, you’d probably have shut them down after six months all the same.
But I'm curious why they bothered to wire up K8s and Kafka at that time. If they did it as a way to study infra with the thought of 'someday launching a service,' then it had its own purpose, and calling that a failure seems a bit ambiguous, doesn't it?
Write the plan in Notion and start by building whatever can be done in a day <- this is the real answer. After I switched to that, I finally got all the way to deployment. The quality is crap, but since it works, people actually used it.
Got a source? "Why did they add this to a service with fewer than 100 users?" That was their own choice, but it seems like too much of a generalization.
I can relate—I've had a similar experience. At first, it's easy to mistake your tech stack for your actual skill. In the end, what remains is a 'working result.' I'm rooting for you to start small next time.
Honestly, the biggest problem is starting without a goal—the stack is basically just an excuse. The moment the word 'someday' comes up, that project is already dead lol
I laughed out loud at my AWS bill lol. Still, realizing it after 6 months is pretty fast. I dragged it out for 2 years before calling it quits.