Quest and airdrop programmes judged on wallet retention

Most quest programmes end with a wallet count and nothing else. Across the last four programmes I have run on Galxe and Zealy, the share of completing wallets still transacting ninety days after the reward vested ranged from 3 to 11 per cent. That spread is the figure a board should be shown, and it is settled before launch by the task set and the Sybil method, not by the size of the reward pool. What you are buying: A quest programme designed backwards from the on-chain behaviour you want ninety days after the claim window shuts. I write the task set, the eligibility rules, the Sybil methodology, the reward curve and the vesting schedule, run the campaign with your community team, and hand back a cohort model that states what a retained wallet cost. How the tasks get built: A task a script can complete produces wallets that behave like scripts. The set is weighted towards actions that cost the participant something other than time: a swap above a floor size, a deposit held for a defined number of days, a Snapshot vote on a live proposal, a Guild.xyz role that requires prior on-chain history, an EAS attestation signed by an existing holder. Follow-and-retweet tasks stay as an entry rung, capped at a fixed share of total XP, and never carry eligibility on their own. Sybil work: Filtering runs twice, once at eligibility and once before the Merkle root is generated. The funding-graph pass clusters addresses by common funder, gas top-up pattern, timing correlation and claim-address reuse, through Dune and Flipside queries I hand over so your analysts can re-run them after I leave. Galxe Passport and Gitcoin Passport stamps are inputs, never the decision. Every exclusion is reviewable: cluster IDs and reasons are published. An accusation of farming you cannot defend in public costs more than the tokens it saved. What gets measured: Day 7, day 30 and day 90 wallet retention by acquisition cohort. Sell-through against the vesting schedule. Cost per wallet still active at day 90. Claimed share of eligible supply, and the size of the unclaimed tail. All of it lands in a Dune dashboard your team owns, with the queries readable rather than buried in a tool I control. Not included: Token design, legal opinion on distribution, exchange listings, market making, Discord moderation, or contract development. I review the distributor and the claim window against the campaign design; I do not write them. Farmed engagement is not bought at any budget. Who this is not for: Teams whose objective is participant count or social reach, since the design here deliberately suppresses both. Teams that need to launch inside three weeks; the eligibility snapshot and the clustering pass take longer than that on their own. Projects with no live product at claim time, because a quest cannot retain a wallet against a testnet. Anyone who wants the Sybil list decided quietly. How the engagement runs: Two weeks of design and snapshot definition, the campaign window itself, then fixed reviews at day 30 and day 90 where I say plainly whether the cohort held. If the day 30 curve is flat, I will recommend cancelling the second reward tranche rather than spending it, and that recommendation has been taken twice. Working language is English. One quest programme at a time.