I stopped brainstorming startup ideas — and started mining Reddit for real problems

Oeljeklaus You — Wed, 29 Apr 2026 14:35:00 +0000

I stopped brainstorming startup ideas — and started mining Reddit for real problems

Like many developers, I used to come up with startup ideas by… guessing.

Sometimes it felt exciting.
Most of the time, it led to building things nobody really needed.

So I changed my approach.

Instead of asking:

“what should I build?”

I started asking:

“what are people already struggling with?”

The shift: from ideas → problems

Reddit turned out to be one of the best sources for this.

Not because of “trends” or hype,
but because people casually describe real frustrations there.

Things like:

“I keep forgetting to follow up on emails”
“there’s no simple way to do X”
“I built this hacky workaround because nothing exists”

These are not ideas.
They’re raw signals.

The problem: it's too noisy

The issue is — Reddit is chaotic.

You’ll find:

noise
one-off complaints
vague frustrations
irrelevant discussions

Manually going through threads is:

slow
inconsistent
hard to scale

My approach: turning discussions into signals

I started experimenting with a simple framework.

Instead of treating every post equally, I look for repeatable signals:

1. Strong problem signals

Clear statements of frustration:

“this is annoying”, “this sucks”, “I hate doing this manually”

2. Ask-for-tool signals

People explicitly looking for solutions:

“is there a tool for this?”

3. Workarounds

Users building their own hacks:

spreadsheets, scripts, manual processes

4. Cross-community repetition

Same problem showing up in different subreddits

Example

Raw discussion:

“tracking follow-ups in Gmail is messy and I always forget”

Signals:

multiple complaints
repeated across productivity + startup communities
users using spreadsheets as workaround

Possible direction:

a lightweight follow-up tracker with persistent reminders

Why this works better (for me)

Compared to brainstorming:

ideas are validated earlier
I avoid building “cool but useless” things
I spot patterns instead of one-off opinions

I built a small tool for myself

To make this process less painful, I built a small internal tool:

scans discussions
extracts signals
groups similar problems
suggests possible directions

It’s still rough, but it helps me:

validate faster
focus on real demand
avoid noise

👉 https://demandradar.vercel.app/

Curious how others approach this

Do you usually:

start from ideas?
or from problems people already talk about?

And if you’ve tried something similar —
what signals do you look for?

Would love to learn how others think about this.

DEV Community: Oeljeklaus You

I stopped brainstorming startup ideas — and started mining Reddit for real problems

I stopped brainstorming startup ideas — and started mining Reddit for real problems

The shift: from ideas → problems

The problem: it's too noisy

My approach: turning discussions into signals

1. Strong problem signals

2. Ask-for-tool signals

3. Workarounds

4. Cross-community repetition

Example

Why this works better (for me)

I built a small tool for myself

Curious how others approach this