How Coconote grows

Web (Stripe) cancel flow: ask why they're canceling, then test three save offers - discount (e.g. 30% off), pause for 3 months (students leaving for summer), and for users cancelling during a free trial 'Do you just need more time?' with 7 extra trial days. Shown in a web view in-app for Stripe subscribers; plans to bring to Apple/Google retention offers.

After iterating one or two onboarding screens at a time (small wins, hard to make big progress), they made one big jump: from ~5 to ~13 screens ending in the paywall. Added screens include login, notification permissions and mostly personalization (what the user is studying). Login was moved to the last screen, after the paywall.

Test roughly doubled onboarding length (to ~15 screens) with more personalization questions and social proof, then paywall; aim is a trial start in the first session.

Removed forced login from the first screen and put it at the end of onboarding, after the paywall; App Store/Play purchases don't need an app account. Requires plumbing to link anonymous purchases to accounts (RevenueCat handles it).

Most customers start on a free trial, and it is hard to win them back once they turn it off. In the cancel flow, besides showing the value they would lose and asking questions, offer to extend the trial by another 7 days.

About 8-9 months after launch, Coconote built pdftobrainrot.com: paste long text and it turns into a Minecraft-parkour 'brainrot' video. Creators on the content team posted videos (text a long message, go to the site, paste it in, reveal the brainrot video) on TikTok and Instagram Reels. The site was a 'Trojan horse' meant to route students into the real study app.

Early (March-May 2024) videos on the @coconoteapp Instagram page used a quick hook plus product demo: e.g. hook 'My mom just changed my college life forever', 2-3 seconds of a lecture hall, then straight to a phone recording the lecture ('all I do is hit record... it listens to the professor, starts taking notes'). Framed Coconote as a solution to a real problem for the target audience. Mostly made by one creator (Joao). Guest says this 'lazy' shock-hook + demo format works while the product category is new and novel.

Launched at $99/year and $19.99/month, pushing the annual plan hard; positioned as premium partly to signal reliability, since users trust the app with exam-critical recordings. Ignored the conventional wisdom that students won't pay.

Built pdftobrainrot.com: turns a long PDF/lecture into a voiceover over background 'brain rot' video (e.g. Minecraft parkour), intended to funnel users toward the main product.

Early price test raised the price to $129 (from $99/yr per the launch pricing); kept it for most of the app's life. Team always had some experiment running.

Each founder put in ~$10K upfront; afterwards reinvested App Store and Stripe revenue into creators; ran at ~50% EBITDA margins every month; used startup credit programs for AI costs early, optimizing costs only later.

Scoped the app down and launched a basic version quickly (April 2024) with a free trial and charging from day one, prioritizing early revenue over polish; founders watched the RevenueCat chart daily as the momentum signal. Accepted early bugs (lost recordings).

One contracted creator ran the main Coconote.app social account, posting short-form videos from early on.

'Navy SEALs, not the Navy': at most 10-12 content creators, all part-time contractors with freedom over what they make. A creator with hundreds of millions of views (Alison) leads a group of contractors, so the lead has the team's respect. The founder interviews every creator (decision within 15-30 minutes). Content coaches (1 per 12 creators) review content and walk creators through metrics. Screening pyramid: wants to win (engaged in Slack, on time, camera on) -> can study other content and predict which videos went viral with view counts hidden -> can create original scripts and trends. Creators split roles by strength (on-camera vs editing) and sit in Slack with engineering.

Moved login/account creation to the last onboarding screen, after the paywall, using Superwall so users can pay before creating an account.

Users can record an entire lecture (an hour-plus of use) before seeing the paywall or starting a trial; the paywall was moved from before the audio recording to after it, so users are invested first.

Reached out to creators via the email in their bio. Target: Gmail address in bio, ~5K-10K followers, good engagement. Skip anyone whose contact is an agency domain ('@sunsetagency.com') - agencies capture the alpha. Rationale: interest-graph feeds make follower count matter less than content quality. Tried ~30 creators in the first two months; kept the ~3 that drove results; creators are part-time contractors.

While experimenting with study games, they built a 2048 game with study questions between moves and promoted the URL 2048buthelpmestudy.com (redirects to Coconote) in a creator video on the brand page. Creator Allie handled hook, script and fast-forwarded sections.

Treat creator recruiting like talent scouting for undiscovered people who see themselves as marketers, not influencers. Sources: Handshake (student job board), Instagram and TikTok DMs, email outreach, referrals, and chance encounters: the founder asked his online Spanish tutor to make content 15 minutes into a lesson; she now makes native Spanish content for LATAM (and some English).

Strategic acquirers emailed from when Coconote was a couple of months old; they took calls with companies they respected and pushed for a ballpark price range after a couple of conversations rather than investing heavily. Once they committed to selling, they hired a bank (Chris Park at Accretive Partners) that handled outreach and conversations through closing, so the lean team kept talking to customers, building and marketing.

Kept the creator team small (5-12; 'Navy SEALs, not the Navy'), founder met each creator 1:1 and shared which creative converts. Measured downfunnel impact on the days a 20-30M-view video went viral. Brief: show the product as the solution to a real problem, or tie it to identity (like Ladder/Strava for fitness); avoid novelty framing, which attracts low willingness-to-pay viewers.

Creators uploaded raw B-roll/unedited footage to a shared Google Drive; an outside clip agency cut it into performance-specific ad creative. Success bar was first-purchase profitability. Neither founder owned the channel more than part-time.

Took an inbound call from Quizlet a few months in; spoke with a handful of companies; once serious, hired investment bank Accretive Partners (Chris Park) to run the process so founders could stay on building. Kept it between the two founders (team and spouse not told) until close; talked to other founders who had been through M&A.

Put some money behind top organic posts like the 18M-view video; kept it minimal after little success.

Hire engineers who flex beyond their main area: iOS engineers also do some backend, web engineers some backend, or one person covers iOS and Android. All apps are native (no cross-platform). AI tooling adds leverage.

Tried five or six tools to measure content creators' success and was disappointed in all of them until Viral App (viral.app), now used to track, manage and optimize the creator program.

The recording and transcribing screens seen in top videos are not real product screens: they are hidden screens only content creators can access, designed to be eye-catching (e.g. a giant rainbow audio wave with 'coconote.ai' in huge text). Creators sit in Slack with engineers and the product team and request these screens.

Asked customers 'How would you describe Coconote?'; answers consistently said it helps them 'never miss a key detail', which became the first App Store preview screen.

Leaned into ADHD from the start in content. The first App Store screenshot says 'Great for ADHD', and screenshot copy like 'never miss a key detail' repeats phrases users wrote in App Store reviews or said in conversations.

Early creator videos used a 'My mom just changed my life' angle. Creator brief: imagine a student and her mom at orientation - the $130 purchase decision goes to the mom. Later de-emphasized, with overlap between audiences.

Users can make one note for free, then must start a free trial; almost no free tier. At launch they also tested trial vs no trial (result not given).