6-Month Side Project Retrospective: I Eventually Gave It Up, but Here’s What I Took Away
Since late last year, I’ve been building a small SaaS product as a side project. I stuck with it for about six months and effectively shut it down last week. I got up to 20 paying users, but it wouldn’t grow beyond that, so I cut my losses when the monthly server costs started exceeding revenue.
At first I thought, ‘I just need to add features,’ but that wasn’t the real problem.
- I built all the features in three weeks. The problem was that nothing happened after that.
- I wrote posts to market it, but I realized too late that what I wrote wasn’t interesting.
- Every customer inquiry ate into my time, and if it doesn’t lead to revenue, it’s just a loss.
Still, I did gain something: I ran the full cycle alone—integrating payments, deploying, and handling incidents—from start to finish. It’s an experience I could never get at a company, so next time I build something, I’ll probably be much faster.
Next time, I plan to ask 10 people before I start building. I’m curious what criteria others use when wrapping up a side project.
6 answers
20 paying users is decent, but the problem was that server costs were exceeding revenue. But isn’t that something you could fix by scaling back the infrastructure a bit? The criteria for cutting losses feels a little disappointing.
I totally agree—I’ve also felt it deeply: marketing is the real bottleneck.
Do you have a source? I’m more curious about how they recruited 20 paying users.
I shut something down similarly, and three months after shutting it down, all I was left with was, “Ah, I should have asked users directly back then.” Next time, I really recommend doing 10 interviews. That’s the fastest way.
Getting payments integrated, shipping, and handling incident response all by yourself over six months is a huge asset in itself. Not many people manage to run that full cycle end to end while holding down a job. I spent three years on mine and eventually gave up, and it seems like you wrapped yours up far more efficiently than I did. Looking forward to your next project.
Well, I'm a bit skeptical of "3 weeks per feature." If payments, auth, and incident response were all really done in three weeks, that just means you built it well—the problem is you didn't validate demand in the first place, not your development speed. And with 20 paying users, you should have architected it so server costs are peanuts; start by reviewing your cost structure.