Building an app portfolio factory: 50 million downloads with a creator army

Executive overview

Most app founders treat TikTok like a lottery. Ryan built a factory instead — a systematic engine for testing thousands of videos, identifying winners, and scaling them with disciplined ad spend.

The core insight: viral is unpredictable, but the system for finding and exploiting virality is not. Ryan's portfolio of 30 apps reached 50 million downloads by combining an in-house creator university, a three-tier ad scaling structure, and aggressive seasonal budget concentration.

A predictable system around unpredictable content is more valuable than any single viral video.

From Facebook UGC to TikTok creator army

• Reflectly (2017) scaled to ~$3–4M ARR on Facebook using micro-influencers before UGC was mainstream
• Authentic, low-production content outperformed polished ads because big brands hadn't caught on yet
• When TikTok launched, Ryan applied the same logic: open season, low competition, platform incentives to spend
• Agencies couldn't crack TikTok internally, so Ryan built a fully in-house creator operation
• The system automated every step: outreach, briefing, concept approval, editing reviews, invoicing, payments
• Target creator profile: 2,000–20,000 followers — proven enough to know the algorithm, small enough to treat it as a business relationship

The in-house creator university

• ~250 active creators from a pool of ~2,000 contacted; 1–2% became reliable viral sources
• Creators were hired as content makers, not influencers — follower count was irrelevant
• Onboarding had three gates: follow the brief, deliver on time, pass a probationary test video
• Ryan's team taught creators the metrics businesses care about: CPI, subscription conversion, hook performance
• Detailed briefs specified trending sounds, hook type, video length (10–15 seconds), shots per second (2+), and required app screen appearances
• A dedicated team member tracked trending TikTok sounds weekly to keep briefs current
• Creators who stayed long-term got regular income and sometimes transitioned to full-time roles

Three-tier ad scaling system

• Level 1: Every video gets a $10–20 test budget as its own ad set — a fair shot at hitting target metrics
• After $20–40 spend, clear signals emerge; underperformers are cut early (sometimes at $5–10)
• Level 2: Graduates move into a higher-spend campaign (100–500+ per day) with fewer videos competing
• Level 3: The top 1% become evergreen campaigns receiving ~90% of total ad spend
• Campaigns at all levels can run simultaneously — once TikTok has learned an audience, don't kill it
• A video can appear at level 1, 2, and 3 concurrently at different spend scales
• Organic posting first is a budget-friendly alternative to paid level-1 testing: post, watch for 5K–10K views, then boost

Seasonal budget concentration

• Health, fitness, and productivity apps have extreme seasonal demand driven by New Year's resolutions
• The first week of January delivers roughly 2x better conversion rates versus the rest of the year
• Ryan allocated ~75% of the entire annual ad budget in January alone
• The December run-up matters: start testing at modest spend (25th–27th Dec), ramp through late December, peak on the optimal weekend
• The day January 1st falls on changes the peak: a Sunday new year outperforms a Thursday new year because people mentally start fresh on a Sunday
• Sunday–Tuesday are the highest-performing ad days for this category year-round; Wednesday–Friday underperform
• The first 2–3 days of each month also see above-average subscription conversion

Content that converts: what actually works

• Organic traction signals ad potential: 5K–10K organic views suggests a video worth boosting
• Comments igniting early (even at low view count) are a stronger signal than view count alone
• Winning formats followed identifiable structures: hook type, editing pace (frames per second), app screen appearances — all tracked and codified
• Viral outliers remain unexplained even after rigorous reverse-engineering (one creator hit 66M views back-to-back with no replicable formula)
• Trend-jacking (familiar voices, trending sounds, known faces) stops the scroll and hands off to TikTok's algorithm to amplify
• Four distribution paths for a single winning video: brand organic post, paid boost, spark ad on creator's account, and ad creative reuse across variants

Portfolio strategy and app acquisition

• After Reflectly, Ryan pivoted to a roll-up model: acquire undermonetised apps, improve paywalls and subscriptions, recoup cost within 12–18 months
• Acquired ~30–35 apps over 2020–2025; the bundle approach spread fixed content costs across multiple apps
• Acquisition criteria: undermonetised, room to improve in-app subscriptions, established user base
• The Mindfulness App (oldest meditation app in Scandinavia) added a high-ARPU segment and opened a B2B pathway
• B2B (corporate wellness contracts) offers six-figure, multi-year deals but requires 6–18 months of sales lead time
• B2B ROI is proven through measurable outcomes (e.g. company-wide stress reduction tracked via in-app tests)

AI agents as the next operating layer (Float)

• Managing 30 apps meant manually aggregating data from App Store Connect, Google Play, RevenueCat, Facebook Ads, TikTok Ads — a full-time operational burden
• Float consolidates all data into one platform with a chat interface for growth insights
• Float 2.0 introduces AI employees: ASO agent, ASA agent, monitoring agent, forecasting agent
• Agents will autonomously handle pricing across 180 countries, custom product page optimisation, ASO, anomaly detection, and fixes — running in the background continuously
• Ryan's thesis: connecting Float to a mobile app will become as standard as connecting Google Analytics to a website