Bootstrapping a Subscription App to 5M MAU and 2X+ Revenue Growth — Bruno Virlet, Genius Scan

Launched in June 2010 as a free app as a 'launch offer', planning to go paid after a few days. After a few days set the price to $0.99 paid upfront. Downloads collapsed immediately, so they panicked and made it free again (later monetized with iAd, a separate paid version and in-app purchases).

Kept the app free and ran Apple iAd banners (high eCPMs at the time). Added a separate paid-upfront version, then, as soon as Apple allowed in-app purchases, a lifetime IAP (Genius Scan+) that removed ads and progressively unlocked advanced features such as quick export to Dropbox.

In 2017, instead of forcing existing premium features into a subscription, added a new subscription-only service, Genius Cloud (document backup and sync). Premium features stayed a one-time purchase. Users accept a recurring fee for storage. In 2020 premium features were folded into the subscription (see grandfathering tactic).

In 2020 renamed Genius Cloud to Genius Scan Ultra and added premium features to it; turned the legacy one-time Genius Scan+ ($9.99) into a subscription tier (priced $9.99/year, or $0.99/month). Two tiers: Plus ~$0.99/month and Ultra ~$3/month, with equivalent yearly prices. Every existing one-time buyer, even those who paid $0.99 in 2010, kept most premium features and can upgrade to Ultra.

Had two subscription tiers since 2020: Genius Scan+ (~$0.99/month) and Genius Scan Ultra (~$3/month), with yearly equivalents. In 2024 removed Plus from sale (existing Plus users kept it) and kept only Ultra. Around the same time the paywall was changed to push the yearly plan. Ultra yearly is now ~$39 (was ~$20-25 in 2020). Price increases are next on the experiment list.

Ultra had monthly and yearly options shown side by side on the paywall. They 'buried' the monthly subscription: still available, but users have to go a bit further into the paywall to find it; yearly is the headline option.

The paywall opens when a free user taps a premium feature (Dropbox export, OCR text recognition). A redesign of the scanned-documents screen moved the OCR button into a drawer. Months later, paywall-display analytics (tracked per entry feature) showed fewer displays; the cause was masked by seasonality guesses. Moving the button back restored displays. They now watch metrics more regularly.

The free version shows a self-promo banner on the main screen (third-party ads were removed) advertising Genius Scan Ultra; tapping opens the paywall. Its text is served from a server so it can be changed remotely, with a hard-coded fallback banner in the app. The server went down (tap counter int overflowed) so the fallback showed; purchases rose. After fixing the server they served both versions (old text vs fallback text/icon) as an A/B test.

The free version was monetized with third-party ads; the best-performing ones were ugly casino-style ads. Founders removed all third-party ads (keeping only an in-house banner for their own premium plan), accepting lost ad revenue for a clean, safe free experience that free users recommend to family and friends.

No onboarding flow and no paywall on first launch, no slow-fading close button. The paywall appears only when users try premium features (export, OCR) or tap the in-app banner. Free users keep a good experience; some convert later and they recommend the app to friends and family.

Email support for every user (free or paid) for 15 years, human answers, no chatbot, no phone. For a long time the two founders answered all tickets; a support hire came ~3 years ago but founders still review support daily and take technical issues. No formal feature backlog: requests heard repeatedly (once, twice, ten times) signal a real need.

Against Apple/Google native scanning and cloud-storage apps' scan modules, invest in scan quality (speed, low-light), document organization and export: e.g. scan and auto-export to a specific Dropbox folder chosen by document name. Targets prosumers and small businesses who research beyond the built-in tool.

Worked through Appfigures-style ASO checklist: app naming and a strong ratings-per-download ratio. Does not buy Apple Search Ads; concluded the apps ranking above them mostly buy App Store ads, boosting download velocity, which drives search rank.

No hiring plan. Hire when the founders feel the pain of too much on their plate: first iOS developer in 2015, a support person ~3 years ago, a product manager this year so the founder can keep coding. Deliberately hire very independent people.

RevenueCat asks sign-ups 'How did you hear about us?' with options plus free-form answers, and the CEO reads every answer daily. Suggested for apps without onboarding: show it to ~5% of users with three big buttons and an easy dismiss.

A daily founder update email lists the 5-10 most important metrics (for apps: total paywall impressions). Each metric is charted as the daily value plus 30-day and 90-day rolling averages: 30-day above 90-day = trending up; daily values catch one-off spikes. No dedicated data analyst needed.