01The problem
Onboarding was killing the funnel before players reached the game. MineAI is a 0→1 AFK mining dApp, but Hotjar showed two of every five users dropping during onboarding, and four of ten test users said the setup was too hard. The five-step native-wallet flow was the wall.
02The thinking
Most arrivals were crypto-native and already ran MetaMask or Phantom. So the fix wasn't a smoother native-wallet setup. It was to drop it: integrate the wallets they already had, two steps instead of five. With the funnel fixed, the real problem became retention. I framed mining as a roster of agents to run, not a yield number to watch.
03Key decisions & tradeoffs
Onboard through wallets players already own
Integrate MetaMask & PhantomnotThe chain's native wallet
ImpactOnboarding drop-off: 2 in 5 → 1 in 5
Agents as your roster
Collectible agents with statsnotA yield dashboard
Strategy over passive yield
Active decisionsnotSet-and-forget mining
One surface, every shell
Shared game statenotDevice-specific apps
04Driving the call
Dropping the native wallet meant getting leadership, including the CEO, to agree not to ship our own. It was a long meeting with real pushback. I won it on the data and the user: GameFi players arrive crypto-savvy and already hold wallets, so a native one only added friction. Past that call, I set the creative direction, built and ran the design system, mentored the other designers, and shaped early acquisition campaigns.
05Outcomes
- Onboarding cut from 5 steps to 2
- Drop-off halved: 2 in 5 → 1 in 5
- 0→1 AFK mining dApp
- One system across web, mobile & dApp
06Reflection
The hardest part was the early complexity. New players still had to get assets like ETH or USDT before they could play, so we leaned on third-party transfer and swap integrations to smooth that over. Getting people to understand the game in the first few minutes was the other risk; we solved it with a guided tutorial and tooltips along the way. Next time I'd design that first-run comprehension up front, not late.
