A Six-Month Retrospective on My After-Work Side Project
Last week, I finally launched a service I'd been building since last winter, putting in one to two hours a day after work. The user numbers are embarrassing to even mention, but just having taken a service all the way through on my own was a big win for me.
At first, I planned about twenty features, but within three weeks I cut them down to five. That was the best decision I made.
I think I asked myself "Should I quit?" about three times along the way.
1. When I was adding login (OAuth hell)
2. When I deployed for the first time (lost five hours to a missing environment variable)
3. When nobody used it
Looking back now, I see a common thread. They were all problems caused by "trying to solve everything alone." When you're stuck, asking the community or just using sample code as-is is much faster.
My next goals are adding payments and SEO. If anyone here has received their first payment through a side project, I'd love to know what verification process you went through.
10 answers
I used Toss Payments for payments, but without a business registration, the card payment review won’t even open. I registered as an individual business under simplified taxation and went ahead with that. For your first payment integration/certification, a domestic PG is just easier.
I'm at a similar point, and the whole "when nobody's using it yet" thing is huge lol. For me, even sending links to people I know makes me cringe so much I just can't do it.
Congrats in advance on your first payment lol. I had to start with business registration and ended up wasting a month on that. Don’t forget to file your mail-order business report too.
I disagree. People often say the payoff of a side project is "the experience of seeing it through to the end," but if you ran a service with an embarrassing number of users, what kind of payoff is that? If you'd just read books or taken courses during those six months, you'd probably have more to show for it. If you're going to write a retrospective, at least disclose the numbers.
Is there a source? Six months is quite long—I’m envious that it made it all the way to deployment.
Before you add payments, first check whether there are actually people who will pay. I built all the features and added payments, but nobody bought anything, which felt like a waste... I recommend just creating a landing page and taking pre-orders. And for SEO, a few blog posts were the most effective.
Six months ㄷㄷ I quit after three weeks—that's impressive.
Well, I think calling it "a problem caused by trying to solve things alone" is a bit of a generalization. OAuth hell isn't something that gets solved by asking the community—in the end, you have no choice but to read the docs. Losing 5 hours there was just a lack of experience; attributing it to being "alone" feels like a stretch. And cutting features from 20 to 5 was the right call, but doing it in 3 weeks isn't something to brag about. If it had been 5 from the start, those 3 weeks wouldn't have been wasted.
I can totally relate to wasting 5 hours on environment variables lol. That’s why I got into the habit of creating a .env.example.
I relate to #3. The hardest part is not being able to tell whether people aren’t using it because of a problem with the service or simply because they don’t know about it. I feel most helpless when there’s no response even after promoting it.