Every founder I know has the same process for finding SaaS ideas. Scroll Reddit. Browse Hacker News. Read Twitter threads. Bookmark interesting complaints. Forget about them. Repeat.
I did this for months. The problem was never finding complaints. People complain about everything. The hard part was knowing which complaints were actually worth building a product around.
The Manual Process
My workflow looked like this:
- Spend 2-3 hours per day reading community posts across Reddit, HN, Stack Overflow, GitHub issues, and Dev.to
- Copy interesting pain points into a spreadsheet
- Try to assess market size, competition, and willingness to pay
- Realize I had no objective way to compare opportunities
- Pick whatever felt exciting in the moment
- Start building
The result? Two products nobody wanted and one that got mild interest but never converted to paying users.
The pattern was always the same. I was selecting ideas based on vibes, not data.
What I Learned After 3 Failed Attempts
After the third failed product, I noticed something. The ideas that had the most potential shared specific traits:
Multiple people described the same problem independently across different platforms. One person saying "I wish X existed" means nothing. Fifty people saying variations of the same thing across Reddit, Twitter, and forums? That is real signal.
People were already paying for bad solutions or building ugly workarounds. If everyone is using spreadsheets or duct-taping three tools together, there is willingness to pay. If the current solution is "just deal with it," you will struggle to charge.
The problem was recurring, not a one-time annoyance. A problem that happens daily drives retention. A problem that happens once a year drives churn.
Search interest was trending upward, not flat or declining. Google Trends takes 30 seconds and tells you if demand is growing or shrinking. Building for a declining market is a trap that looks like a good idea until 6 months in.
The ideas I picked based on gut feeling had none of these signals. They sounded cool but did not pass any of these filters.
Why Manual Validation Does Not Scale
The manual approach works if you are validating one idea. It breaks when you want to systematically find the best opportunity across thousands of community posts.
Reading 50 Reddit threads a day gives you a tiny sample of the problems people are talking about. You miss entire communities. You miss patterns that only show up when you look at data from 9 different sources simultaneously. You miss the signal because the noise is overwhelming.
And the biggest problem with manual validation is confirmation bias. Once you get excited about an idea, you start filtering for evidence that supports it and ignoring evidence that contradicts it. You need a system that scores opportunities objectively before you get emotionally attached.
Automating the Process
I built Drok AI to do what I was doing manually but at scale. It scans online communities for real pain points, then runs each one through a three-stage AI pipeline:
Stage 1: Validation. Is this a real SaaS opportunity or just someone venting? The pipeline filters out noise, feature requests for existing products, and problems that are not viable businesses. Only ideas that score above an 80% viability threshold move forward.
Stage 2: Deep Analysis. Market size, competition, feature requirements, pricing potential, and competitor mapping. This is the step that would take hours of manual research per idea. The AI does it in seconds.
Stage 3: Trend Enrichment. Is demand growing or shrinking? What related queries are people searching for? What does the historical interest curve look like? This is where Google Trends data gets layered in to show whether the opportunity has momentum.
Every idea gets scored on viability. Instead of a messy spreadsheet full of bookmarks, you get a ranked feed of opportunities backed by actual data.
The Self-Validation Feature
The part I use the most is actually the self-validation tool. If you already have an idea, you can run it through the same pipeline at SaaSBrowser.ai and get an honest assessment before writing any code.
It scores your idea across the same dimensions: severity of the problem, frequency, willingness to pay, competition, and market trends. No confirmation bias. No friends telling you "yeah that sounds cool." Just data.
This feature alone would have saved me from building those two products nobody wanted.
What I Would Tell My Past Self
Stop reading Reddit for 3 hours a day hoping to stumble on the right idea. Either systematize the research or accept that you are gambling with your time.
The best SaaS products solve problems people are already vocal about. The signal is out there. You just need a better way to find it.
If you are in the ideation phase, check out SaaSBrowser.ai. Daily Insights are free once you sign up.
Dylan is the founder of Drok AI, an AI-powered platform that discovers and validates SaaS opportunities from real community pain points.
Top comments (0)