A crypto quest platform is hosted infrastructure for running task campaigns: a project publishes a list of actions, the platform verifies completion, and participants earn a credential or a reward. Three platforms hold most of the volume in 2026, and they are not interchangeable. Galxe reports more than 21 million unique addresses having taken part in quests across roughly 4,900 partner programmes. Zealy runs at about 700,000 monthly active users. Layer3 is the smallest of the three and the most on-chain.
The reason to care about the difference is that each platform attracts a different participant, and the wrong choice produces a large number that means nothing. Arbitrum found roughly 149,000 sybil addresses in its distribution, capturing an estimated 21.8% of the tokens. That is the failure mode this channel is prone to, and it is a design problem rather than a platform defect.
Which quest platform should you use?
| Platform | Scale | Best for | Weakness |
|---|---|---|---|
| Galxe | 21M+ unique addresses, ~4,900 partners | Top-of-funnel reach, on-chain credentialing, large mints | Volume attracts farmers; verification design is on you |
| Zealy | ~700,000 monthly active users | Discord-era community grind, social raids, content tasks | Social tasks are the easiest to fake |
| Layer3 | Smallest of the three | Protocol-level retention, education plus on-chain depth | Smaller audience, slower to fill a campaign |
| Intract, QuestN, TaskOn | Smaller regional and niche audiences | Cheap testing, specific geographies | Thin verification, uneven audience quality |
| Guild.xyz | Access control rather than quests | Gating a Discord or app by wallet or credential | Not a campaign tool on its own |
One naming point that causes confusion: Crew3 is not a separate platform. It rebranded to Zealy in April 2023, and material still comparing the two is describing one product under both names.
What does each platform actually do well?
Galxe
Galxe is the reach option and the credentialing option. Its scale is the reason to use it and the reason to be careful with it: a campaign that opens to the platform’s reported 21 million addresses without a verification tier will be found by people who complete quests for a living. Where it earns its place is on-chain credentials that persist after the campaign, which lets you segment later rather than starting from zero each time.
Zealy
Zealy is built for the community phase rather than the launch phase. It suits sustained social activity, content submissions and the daily rhythm of a Discord that needs a reason to open. The trade-off is that social tasks are the cheapest in the world to fake, so a Zealy board without review produces follower counts and nothing else.
Layer3
Layer3 is the option to reach for when the goal is a user who understands the protocol. Its quests tend to combine reading with an on-chain action, which filters harder than a follow button and produces fewer, better participants. If the brief is retention rather than awareness, this is usually the right starting point.
Why do quest campaigns attract farmers?
Because the reward is the product. Quest farming means completing the minimum required actions purely to qualify for a token or NFT, with no interest in what the project does. It is recognisable after the fact by a high claim-and-sell rate and near-zero engagement afterwards, and it is not a fringe problem: the Arbitrum distribution lost an estimated 21.8% to roughly 149,000 sybil addresses.
The mistake behind most farmed campaigns is measuring the wrong thing while the campaign runs. Completion count rises smoothly whether the participants are real or not, so a dashboard showing tasks completed will look healthy right up to the moment the token unlocks.
How do you run a quest campaign that keeps people?
Verification tier first, reward second. In our own task campaigns we run two levels: a basic tier that confirms the action was genuinely performed, and an advanced tier that adds participant KYC. The second costs more per participant and removes most of the population that would have sold immediately. Campaigns run against a floor of 5,000 verified participants and a ceiling of 10,000 a month, which is deliberate: a slower campaign that can be checked beats a fast one that cannot.
Three 2026 campaigns built this way moved member counts from 35,000 to 47,000, from 130,000 to 141,000 and from 40,000 to 45,000. The number we watch afterwards is not the peak but the figure on day 30, because that is the first point at which a campaign has to survive without a reward attached to it.
- Gate the reward behind an action the product actually needs. A wallet connection and a first transaction filter far harder than a follow.
- Stage the reward. Paying part on completion and part after a holding period removes the claim-and-sell cohort without touching genuine users.
- Cap participation. An uncapped campaign is an invitation to farms, and the marginal participant above the cap is almost never the one you wanted.
- Report day 30, not day 1. Peak members is a vanity figure. Retained members is the campaign result.
What should a quest campaign report show?
A useful report separates the people from the actions. Total task completions belong in it, but alone they say nothing. The figures that carry meaning are verified participants against total participants, wallet connections, first transactions, and the retained share at day 30. If a supplier reports only completions and total reach, they are reporting the number that is easiest to inflate.
We run these campaigns with a verification tier rather than an open form: how we run task campaigns.
How to vet crypto quest platforms before launch
Vetting starts before the sales call. Open the platform’s archive of finished campaigns rather than the ones it is promoting, and read how the tasks were worded, what completion actually required, and how the project answered complaints in its own channel once rewards went out. A platform whose closed campaigns are all follows and reposts will hand you the same campaign under a different name, whatever the pitch says it can do.
Then run the flow yourself with an empty wallet. Complete two tasks and submit one deliberately sloppy answer, then watch what the checker accepts. If a screenshot passes where an on-chain action was promised, you have found the ceiling of that platform’s verification, and no campaign design above that ceiling will repair the data you get back. Note how long manual review takes, because that waiting time becomes your support queue during the campaign.
Ask the billing questions in writing. Whether you pay per verified participant or per campaign slot changes every decision that follows, and so does the answer to what happens to participants rejected at review, because some suppliers bill for them anyway. Ask who holds the reward pool, when it has to be funded, who tops it up if it empties, and what happens to rewards nobody claims. Vague answers at this stage reliably predict vague reporting later.
Scoping tasks before the budget
Write the task list before deciding what to spend, not after. Each task should map to a step in your own product that you would want a real user to take this month. If you cannot name the internal number a task moves, it is decoration, and it will attract participants who are also decoration. Four or five tasks is usually enough, because long boards raise the cost of completion without raising the quality of the people completing it.
Order matters more than length. Put the cheapest social action first so the funnel fills, place the wallet connection or first transaction in the middle where intent has already been shown, and keep the last task light so genuine participants do not abandon a board they have half finished. Watch the drop-off between steps while the campaign runs. A cliff at one step is almost always wording or a broken link rather than reluctance.
Contract terms worth pinning down
Put the verification tier into the statement of work, in the same sentence as the price. If a supplier quotes against a basic tier and delivers unreviewed completions, the invoice will still look correct and you will have no grounds to argue. Specify what evidence arrives per participant and in what format, then set the date it is due. A file of addresses with completion timestamps is the minimum that lets you check anything at all after the fact.
Name the caps in the contract as well. A floor and a ceiling on participants per month stops a supplier front-loading delivery into week one and leaving you with a campaign that peaked before the product was ready for the traffic. Agree in advance what happens when the floor is missed, whether that is a refund, an extension or a rollover into the next month. Settling that question while a campaign is failing never ends well.
Running a small paid trial first
Before committing a quarter of the budget, buy the smallest campaign the platform will sell and treat it as measurement rather than marketing. A trial answers questions the sales deck cannot: how fast the board fills, how many completions survive review, how many participants ever reach the product, and how responsive the account manager is on a normal working day. Budget for the trial to produce nothing except an answer.
Compare the trial against a channel you already trust rather than against the supplier’s own benchmark. If your existing community produces better returning users at similar cost, the quest layer is adding administration rather than people. Run the comparison on the same window and the same definition of a retained participant, otherwise the two numbers are not describing the same thing and the decision they support will be wrong.
When a quest campaign goes wrong
Most campaigns fail quietly rather than dramatically. The signal is a completion curve that keeps climbing while wallet connections flatten, which means the board is being cleared by people who never intended to touch the product. Pause rather than cancel. Cancelling mid-campaign burns the participants who were genuine and hands the supplier an argument about delivery, while a pause lets you rewrite the task list and restart against the same budget.
Reading the participant list afterwards
Export the address list and sort it by first activity. A cohort funded within minutes of each other from the same source, completing the same tasks in the same order at the same pace, is a farm, and it shows up in the timestamps long before it shows up in the totals. Strip those addresses out before reporting anything, and keep the excluded list so the next campaign starts with it already loaded.
Then look at the other end of the distribution. Participants who did more than the minimum, came back without a prompt, or asked a question in your channel are the real output of the campaign. They are usually a small share of the total, and they are the group worth a follow-up message or an invitation into whatever you run next. Reporting only the aggregate hides exactly the people the campaign was supposed to find.
When rewards fail to reach people
Reward delivery breaks more often than campaigns do. A claim contract runs out of funds, a token address is entered with the wrong decimals, or a participant qualifies on one address and claims from another. Decide before launch who answers those messages and where, because a complaint posted publicly and ignored for two days costs more reputation than the reward was worth. Keep a short written answer ready for the common cases.
Fix the ledger rather than the individual case. When a delivery problem appears, freeze claiming, reconcile the qualified list against what actually paid out, then reopen with the corrected list published. Paying quietly on request rewards whoever complains loudest and leaves the honest participants unpaid, which is the version of this failure people remember and repeat for months afterwards.
Tracking quest traffic end to end
Tag every link leaving the quest board so traffic arrives in your analytics with a campaign name, a platform name and a task identifier attached. Without the task identifier you cannot tell which step produced the people who stayed, and that is the only finding worth carrying into the next campaign. Check the tags on a live board yourself, because platforms rewrite and strip parameters more often than their documentation admits.
Match the on-chain side separately. Addresses that connected during the campaign window belong in their own segment, and the number to watch is how many of them transact again once the reward is paid and no further task exists. Day 30 is the checkpoint this page argues for throughout, and it only exists if you recorded who arrived and from which task they came.
Two edge cases deserve a plan in advance. A campaign that fills far faster than forecast is usually being farmed rather than loved, so treat early completion as a reason to inspect the list rather than to celebrate. A campaign that will not fill at all is often a timing problem: the audience a platform reaches sits in hours when your channel is asleep, and moving the announcement window fixes more of it than raising the reward.
Quests often feed a mint raffle; our guide to NFT raffle strategies covers the draw itself and sybil filtering.
Frequently Asked Questions
Which quest platform has the most users?
Galxe, by a wide margin. It reports over 21 million unique addresses having participated in quests across roughly 4,900 partner programmes. Zealy runs at around 700,000 monthly active users, and Layer3 is smaller than both.
Is Crew3 the same as Zealy?
Yes. Crew3 rebranded to Zealy in April 2023. Any comparison presenting them as two competing platforms is out of date.
How much of an airdrop typically goes to farmers?
It varies by design, but the scale can be severe. Arbitrum identified roughly 149,000 sybil addresses in its distribution, estimated to have captured about 21.8% of the tokens.
Which platform is best for retention rather than reach?
Layer3, because its quests usually pair education with an on-chain action, which filters more effectively than social tasks. Galxe is the better choice when the goal is reach and persistent credentials.