They Said They Loved It. But Would They Pay?
100 people signed up. You're excited. But do they actually need this, or were they just curious? Here's how to find the truth.

1. The Trap of 100 Signups
Don't get drunk on numbers. "100 signups" only matters when you convert it to "12 people who completed a core action."
The central question of Stage 2 is simple: "Can these people live without this service, or did they just look around?"
2. User Interviews — Five People Is Enough
Surveys lie a lot. 80% of people who check "this seems useful" never actually use the product. A 30-minute 1:1 conversation is far more powerful.
Design your interviews around past behavior, not hypotheticals. Don't ask "Would you use this?" Ask "How did you manage your appointment notebook last week?"
Other option: Five-second usability test — if interviews feel heavy, show your screen for five seconds and ask "What is this and where would you click?" to get quick directional signal.
3. Reading Signals — Good vs Bad
- They come back on their own
- They mention recommending it to a friend
- "I can't run my shop without this now"
- Feature requests are specific ("2-week reminder")
- 7-day retention above 30%
- They visited once and never returned
- They praise it but don't actually use it
- "I'd just keep doing it by hand"
- Requests are vague ("make the AI better")
- 7-day retention below 10%
4. The Courage to Kill — Knowing When to Stop
If bad signals outnumber good ones, you need the courage to kill the project. It's hard because of sunk cost. But killing well is a skill.
Most projects should die at this stage. That's not failure — it's learning. The real value of vibe coding is that you can build cheap and kill cheap.
5. The Stage 2 Mindset
J called five salon owners for 30 minutes each. Three said "appointment notebooks are a pain." One said "I already use another app." One said "we're fine with pen and paper."
When J asked "What would you do without this?" three of them said "I'd just write it in my notebook." That's the answer. The notebook is inconvenient, but not inconvenient enough to force a switch. J decided to add a "2-week advance reminder" feature — something no notebook can do.
Stage 2 is a war against self-comfort. Drop the feeling that "people like what I built" and focus only on "do people actually use this?" Pass that test, and you can move forward.
About the Author

Jaehee Song
Enterprise data platform architect with 20+ years of experience building data systems for Fortune 500 companies. AI development educator who has taught vibe coding and AI development to hundreds of students. Founder of Seattle Partners, helping Korean technology startups navigate the US market.