How Erly grows

Built the MVP (a simple accountability app to log your wake-up against a goal set the night before) in ~19 days from first commit to App Store, deliberately without the alarm feature. Treated views, not building, as the unknown, so started marketing immediately: first influencer video went live 7 days after launch. Purchases from that one video counted as validation, then scaled up.

Found TikTok and Instagram creators already making day-in-the-life or lifestyle videos where the app fits naturally (they wake up on camera) and sent a ton of DMs, plus mass outreach emails via Instantly. Early reply rates were very low. Pitch to creators: keep making your usual videos and get paid for them. Example: 'POV: Day in the life of a college student' showing the alarm going off, doing push-ups, the app analyzing and a green check 'alarm turned off'. Well-shown videos draw 'What's the app name?' comments; the creator replies 'Erly on the App Store' and usually pins it on Instagram. Tools: Growee for influencer tracking and payments, Sideshift for UGC creator sourcing/payments, Cal for meetings, Agree for contracts, Google Sheets.

Typical deal: 2-4 videos per month with a view guarantee, so payment tracks views delivered; effective rate around $2-3 CPM. Cap what you pay, then judge the return; a video reaches peak views in about a day. Returns follow a power law: most videos do okay, occasional 1M/2M/5M-view videos make the money. Expect to lose money on some individual creators without knowing which.

The accountability MVP wasn't viral enough for influencer marketing, so Jake spent a day walking and writing out possible directions and chose an alarm feature: 'push-ups to turn off your alarm'. Timed it with iOS 26's AlarmKit API, which lets third-party apps use the native iPhone alarm (earlier alarm apps relied on unreliable sound workarounds). The app films you and checks you actually did the push-ups. The one-liner is pushed through all marketing. His rule: solve a painful problem simply, and if the idea isn't inherently viral, add a novel aspect that is.

Keeps it simple: tracks views on every influencer video with historical data (Google Sheets, Growee) and compares against conversions (Mixpanel/Superwall). When a view spike coincides with a conversion spike, credits that video. Treats the channel as a black box judged on total spend in vs revenue out, accepting that some individual creators lose money unidentified.