Zero to 1.5M MAUs in 90 Days
I ran growth for Ninety Eight’s mobile gaming platform, June 2023 to April 2025. First 90 days: zero to 1.5 million monthly active users.
Ninety Eight had a crypto finance platform, Coin98, and wanted a gaming ecosystem on top. My pitch to studios: bring your game, I bring the users. But there were no users, and no studio joins an empty platform. One quarter, $50K a month, to manufacture the scale that closes the deals.
The loop did the work
The engine was a referral mechanic inside the game, not paid ads. Players recruited friends onto their teams, and team composition drove in-game performance, so recruiting was a real edge. Good formations got shared and argued over, all free. Play, build a team, share the formation, friends join, better team, share again. It hit a 1.2x viral coefficient: every 10 users brought 12 more, and paid became a top-up.
Onboarding was the leak
Day-1 activation launched at 42%. Over half churned before hitting real value: onboarding took 15 minutes across wallet setup, KYC-lite, tutorial, and first game. I ran multi-wave A/B tests on the first session and gutted it. Wallet setup dropped to two taps over Telegram login, no separate crypto wallet. KYC-lite moved to first withdrawal. The tutorial became a guided first match that taught by playing. Time-to-value went 15 minutes to 3, activation 42% to 60%, retention 45% across the full 1.5M base.
Spending the $50K
The $50K ran across 10+ gaming studio partnerships, each promoting its own game on Cyborg while I owned the platform’s spend. No big bets. I reallocated weekly off CPI and retention data, per channel, per game, per geo, killing losers and doubling winners. CPI stayed at or below $0.20 across the entire 90-day ramp. By 1.5M MAUs the loop out-minted paid, so the budget rode on top of organic instead of funding it. Cumulative paid media across Garena FO4 and Cyborg is $1M+, a conservative floor from daily operating norms at both.
The Numbers
| Metric | Before | After | Timeline |
|---|---|---|---|
| MAUs | 0 | 1,500,000 | 90 days |
| Viral coefficient | — | 1.2x | Sustained |
| Day-1 activation | 42% | 60% (+18pp) | A/B tested over 6 weeks |
| Retention | — | 45% | Sustained at scale |
| Time-to-value | 15 min | 3 min | Re-engineered onboarding |
| CPI | — | $0.20 | On $50K/month budget |
| Transaction volume | baseline | +30% | Incentive campaigns |
| Studio partnerships | 0 | 10+ | Ongoing |
What compounded
Without the viral loop, paid would have remained the primary engine. Retention was never separate from acquisition: at 1.2x, every retained user is also an acquisition channel. Moving activation from 42% to 60% kept more users in the system, and the loop fed each into the next wave. Fix retention first, then turn on paid.