Ask HN: How did you find a side project idea that worked?
Hi HN,
Over the past year I’ve started and abandoned quite a few side projects.
Most of them followed a similar pattern:
- I’d see an idea (often from Twitter, Reddit, or other builders)
- It would feel promising at first
- But after digging a bit, it either turned out to be too crowded, or I couldn’t tell if anyone actually needed it
The hardest part for me isn’t building — it’s deciding what to build with some level of confidence.
I’ve tried a few approaches:
- Building from personal pain points (sometimes too niche)
- Looking at what others are launching (often already saturated)
- Checking search trends / keywords (but hard to tell signal vs noise)
It still feels very hit-or-miss.
For those of you who’ve built something that got real users:
- How did that idea originally come together?
- At what point did you feel “this might actually work”?
- Was there any concrete signal, or was it mostly intuition?
Lately I’ve been thinking about whether there’s a more systematic way to spot early demand signals — especially from things like search behavior — but I’m not sure if that’s a useful direction or just another rabbit hole.
Curious how others here think about this.
1 comments
A reliable way to validate interest is to look for specific frustration patterns in Reddit comments rather than just broad topics. You want to search for phrases like "how do I" or "is there a way to" within niche subreddits related to your industry.
Negative sentiment is usually a stronger signal than positive feedback. If people are complaining about a specific feature in a popular tool, they're likely ready to switch to a simpler alternative that solves that one problem.
I built Reddinbox to help with this exact research phase. It uses AI to scan these communities and pull out high-intent signals and pain points so you don't have to manually read through thousands of threads to find a gap in the market :)
Negative sentiment is usually a stronger signal than positive feedback. If people are complaining about a specific feature in a popular tool, they're likely ready to switch to a simpler alternative that solves that one problem.
I built Reddinbox to help with this exact research phase. It uses AI to scan these communities and pull out high-intent signals and pain points so you don't have to manually read through thousands of threads to find a gap in the market :)