How Canary grows

Posted a video on his personal account of himself singing in French with the lyrics and translation; it got a lot of engagement, mostly shares and new follows. Instead of relying on his personal audience, he created a separate page just for this content (not about him). It grew faster than his personal page, to 200,000 followers in under a year. He treated that as social proof that people want to learn with songs and then built an MVP. He says you can also build an audience from scratch to test an idea.

When a piece of content performs very well organically, boost it as a Meta ad. Test creative, audience and caption variations, scale the winning ads and kill the losers. Full loop: post a lot, find a winner, make variations, boost the winner with paid social, kill the loser, look at the numbers, repeat.

Freemium app with a premium subscription at $10/month or $79/year. They push the yearly plan. Uses RevenueCat for subscriptions and paywall A/B testing (specific paywall tests not described).

Hired a developer (he can't code) and shipped the first MVP in about a month, even though it was 'really ugly and really bad'. Mindset: launch something and get feedback as soon as possible, then improve; 'ready is always a work in progress'.

After starting with a hired developer, about four months before recording he began vibe coding the app himself with Claude ($200/month). Lets him build features quickly, act on feedback and change things in the app directly.

Early on, posted on every platform with different content and formats ('A-B-C-D-E testing'). Studied other apps and listened to founder podcasts (Starter Story) every day at the gym, taking notes on things to test. Routine: every day learn something, ship something, test something.

Reached out to past Starter Story guests, including Nicole, and paid her for a consulting session to help improve the onboarding and the product.

Choose your platforms so you're not overwhelmed. His main ones were Instagram and TikTok; Reels were what worked for him, so he doubled down on Instagram. Post a lot, find winners, make variations of them.

Added Manychat (about $30/month) for social media automation to convert engagement on his posts into traffic to the app. Specific triggers or messages not described.

Spend time in content from a niche you enjoy and watch what people complain about or struggle with. Many of his followers used Duolingo but couldn't say anything in French when he asked; the only thing they remembered was a song lyric. He concluded people remember lyrics and built a language app around songs.

Tried Google Ads after hearing a Starter Story founder talk about them. No setup or results given.