How a disposable camera app hit $20K/month in 83 days

Executive overview

Most founders build first and validate later — and waste months on the wrong thing. Brian Shin flipped this: he defined a commitment metric and refused to write a single line of production code until he hit it.

The result was $20K/month within 83 days of launch. The playbook works because commitment (not just interest) filters out polite enthusiasm from real demand.

Validation is not about building — it's about getting strangers to commit before you build.

The app: Once

• Digital disposable camera for group events — weddings, birthday parties, corporate events
• Guests join via invite code, take photos; all images reveal at a set time like a real disposable camera
• Pricing by guest count: $2 for 10 guests up to $50 for 150 guests
• Launched December 2025; 10,000–12,000 weekly active users by month three
• ~700 events scheduled in March 2026

The commitment metric framework

• Commitment metric: a pre-launch signal that proves real demand — not payment, but a meaningful action
• For Once, the metric was: 10 event hosts who committed to an actual date and would use the product at a real event
• A commitment tied to a social event (with friends and family present) was treated as near-equivalent to payment
• Set a deadline for yourself alongside the customer target — makes validation a finite, focused sprint

The 5-step validation playbook

1. Define your commitment metric — pick a specific number and date; hold the line until you hit it
2. Exhaust your personal network — open every social platform, scan your contacts, filter for people with upcoming events; apply the mom test to stay honest
3. Build a quick mockup — two to three days maximum; scrappy is fine; the goal is something you can show, not something that works perfectly
4. Go where your users are — Reddit, TikTok, Instagram, YouTube; pick one channel and go deep; cold message at scale
5. Hit the number before building — Brian sent 250–300 cold messages, got 15 responses, locked in 12 events; that was the green light

Building and tech stack

• First version was a web app built in one to two weeks, tested at a Halloween party — broke repeatedly but validated the core idea
• After web validation, scrapped v1 entirely and rebuilt as a mobile app
• Design is done without AI; "consumer apps are a craft" — taste matters
• Dev: Claude Code (often running multiple instances via Conductor on separate work trees)
• Backend and database: Supabase
• Design: Figma

Key mindset shifts

• AI tools make building easy — which makes it tempting to skip validation; resist this
• Cold outreach feels uncomfortable; if you haven't been banned from a platform twice, you haven't pushed hard enough
• Stop overthinking distribution and launch fast — assumptions rarely survive contact with real users
• Bootstrap path offers control and lower stakes than venture-backed startups; neither is wrong, but they require different mindsets