Why I Built a Tool to Hunt Problems (Not Products)

For months I was building features nobody asked for. So I flipped it: instead of building first and hunting for users after, I wrote a pipeline that finds people already describing the exact problem my product would solve. It isn't clever—just a practical, scrappy automation that saves hours of manual scrolling. Here’s the hands-on walkthrough.

Why most discovery methods are broken

• Surveys are biased by what people think they want.
• Keyword research shows volume, not actual frustration.
• Asking for feedback requires an audience you don't have yet.

My tool listens to conversations where the pain is already raw: Reddit, Stack Overflow comments, and Twitter search. The output is a spreadsheet of real humans actively complaining, with link, context, and engagement score.

Step 1 – Build your problem signal list
Don't