If I restarted Forge Regex this week, I’d cut half the distribution and none of the tester
Hey IH —
A month of shipping Forge Regex solo. Hypothetical: one week of runway, restart today.
Keep
- Describe → regex → live highlight on the same page
- Editable pattern + refine
- Free daily cap that doesn’t blank out testing when you hit zero
- Paddle checkout that matches the listed $9 (I already burned time on a $1 test Price ID)
Cut / delay
- Dual-market payment experiments before overseas demand exists
- Daily auto-comments on IH (habit, not a channel)
- Cloud history as the main Pro story, until anyone uses local Recent
- More launch-style posts until must-match / must-not examples exist
Principle I actually believe now
The first 30 seconds have to prove the regex works on your text. Distribution without that just produces nice replies and an empty generate log.
Ask
What’s the most expensive distraction you shipped before the core loop was obvious?
The core loop is clearly the priority. What are testers consistently failing to accomplish in those first 30 seconds before the distribution or Pro questions even matter?
That zero-row history table you mentioned last week leaves two very different possibilities: nobody really tried the workflow, or people tried it and had no reason to come back. Those point to different problems. What happens after someone runs one regex on their own text: do they save or share the result, or just leave?
The tester is the part that actually teaches you what to keep. Half the directory and board list usually just burns a week and the generate log stays empty. Cutting distribution first is the move most people refuse because it feels like less effort means less chance, then they just ship noise everywhere.