Growth · Web3 · Community
Scaling a Web3 ecosystem to 1.4 million users: what worked
The acquisition, retention and incentive systems behind the G-Agents AI ecosystem's growth to 1.4 million users on TON, and what I would do differently.
When I joined G-Agents AI (then Gamety) in May 2024, the TON ecosystem was heating up fast. The project had technology but no audience. My job was to lead the community work that would turn a product into an ecosystem.
Telegram first
In Web3, community lives on Telegram. We started with a clear brand voice: education first, no hype. The first 30 days were about content quality: AMAs, explainer threads, ecosystem updates and honest roadmap discussion.
The Telegram group reached 50,000 members in the first two months. The playbook was simple: three to four posts a day, fast answers to questions, a weekly community call, and no FUD left unanswered.
The Zealy campaigns
We set up Zealy (formerly Crew3) campaigns that rewarded real behaviour: following the project, engaging with content, inviting friends and completing on-chain tasks. The point was to make each task useful to the person doing it, rather than hollow engagement farming.
One sprint brought in about 29,000 new members. That took the task design, the reward structure and the timing working together.
KOLs: quality over volume
Many Web3 projects spend heavily on large KOLs and hope the followers convert. We went the other way: smaller, niche KOLs with 10,000 to 100,000 followers and engaged crypto-native audiences. The CPM was higher, but conversion was four to five times better.
The Discord layer
Telegram drives acquisition. Discord drives retention. We built Discord around role-based channels, on-chain verification and an active moderation team, and reached 45,000 members in three months with strong daily activity.
Results, and what I would change
About 1.4 million ecosystem users in six months. 401,000 Telegram members, 229,000 followers on X and 45,000 on Discord. If I did it again, I would build the analytics dashboard sooner. We tracked everything by hand for the first two months, and it slowed every iteration.
Community growth in Web3 is rarely about spending more. It comes from the system you build: the feedback loops, the incentives, and the quality of every conversation your team has with potential users.