Community building or advertising: which does a web3 project actually need?
Community is not a cheaper substitute for advertising. It is a permanent staffing cost with a payback measured in quarters, and it is very poor at the one thing advertising does well, which is reaching people who have never heard of you. Web3 leans on community anyway, largely because the paid channels are closed rather than because community is superior. Understanding which constraint you are working under makes the allocation obvious.
Can a community replace paid advertising?
No, because they solve different problems. Advertising buys first contact with people who do not know you exist, at a price you can raise or lower on demand. A community converts, retains and defends interest that already exists. Neither substitutes for the other, and a project with no top of funnel will find that a busy Discord makes no difference to the numbers.
The mistake is treating community as free reach. It is owned, which is not the same as free: you pay for it in moderators, timezone coverage, security and the attention of senior people, and it takes months to build to a size where it does anything. Advertising is rented and instant. The cost structures are opposite, and so are the risks.
Where community genuinely outperforms is durability. Ad performance stops the day you stop paying. A community that has learned to answer its own questions keeps working, which is why the sensible framing is not either or but which stage of the funnel each is being asked to fix.
Why does web3 lean on community so heavily?
Mostly by necessity. The major paid platforms restrict or prohibit advertising for cryptoassets, exchanges, wallets and in some cases NFTs, and the exemptions require certification and are limited by country. A project that cannot buy reach organises the reach it can get, and that means chat platforms, founder accounts and holders.
There is a second, better reason. Holders have a financial or membership stake in the thing succeeding, which no ad audience has. That produces genuine distribution: people who explain your product to others because they benefit if more people understand it. It also produces the failure mode, which is an audience whose enthusiasm is proportional to price.
Both reasons should be stated plainly when planning, because they lead to different designs. If you are in community because ads are closed, you are substituting labour for media and should budget accordingly. If you are in community because holders genuinely evangelise, the job is to give them something to say that is not about price.
What does each channel actually buy?
Four options are realistically available to a web3 project, and they buy different things on different timescales. The table is worth reading as a sequencing tool rather than a menu, because most projects need more than one and in a particular order.
| Channel | What it buys | Time to first result | What breaks it |
|---|---|---|---|
| Owned community | Retention, support, defence during drawdowns, word of mouth | Months | Understaffed moderation, or nothing to discuss but price |
| Paid advertising | First contact with people who have never heard of you | Days | Platform policy, certification requirements, account bans |
| Paid creators and KOLs | Borrowed trust and reach inside an existing audience | Weeks | Undisclosed payment, which is both unlawful and obvious |
| Incentives and airdrops | Wallet count and short-term activity | Immediate | Farming: the wallets leave the moment the incentive ends |
| Earned press and research | Credibility with people who distrust the category | Months | Having nothing technically interesting to say |
What does running a community actually cost?
It is a headcount line, not a media line. A public chat that is open around the clock needs moderators around the clock, because the attack surface is continuous: impersonation of admins, phishing links in direct messages, compromised webhooks, fake support accounts responding faster than your real ones. A community left unmoderated overnight is worse than no community, because it becomes a distribution channel for scams that carry your name.
Add to that the senior time. The parts of community work that cannot be delegated are the parts that matter: answering hard questions about the roadmap, admitting a delay, explaining a decision that lost people money. Delegating those to a junior moderator with a script is the fastest way to lose the room.
Budget it as an ongoing operational function with a security component and shift coverage, and the decision between community and advertising becomes a straightforward comparison. Budget it as a campaign and it will be quietly abandoned within two quarters, which leaves an empty channel that signals more than silence would have.
How do you tell whether your community is doing any work?
Stop posting for three days and watch what happens. If member to member conversation continues, questions get answered by people who do not work for you, and nobody assumes the project is dead, you have a community. If the channel goes silent, you have an audience waiting for announcements, which is a mailing list with worse deliverability.
The second test is to read the last two hundred messages and count how many are about price. A room that only discusses price is not a marketing asset, because its members cannot explain your product to anyone. They will also disappear on the first sustained drawdown, and they will take the mood with them.
Both of these are rough tests you can run this afternoon. Proper measurement of concentration, retention and holder behaviour is a separate discipline and worth doing before you commit real money to the channel.
So how should the budget actually split?
Fund the community as a permanent function at whatever level you can sustain for a year, and treat everything else as variable. That order matters, because an understaffed permanent function fails in a way that damages you, whereas an underfunded campaign simply does less.
Then spend the variable budget where the funnel is actually blocked. If people arrive and do not convert, the problem is the claim flow and the proposition, and more reach makes it worse by scaling the leak. If nobody arrives at all, no amount of community management fixes it and you need bought or borrowed reach.
One rule holds regardless of split: do not buy incentives to solve an awareness problem. Airdrops and reward campaigns produce wallets, and wallets are not people. If the metric that improves is wallet count while retention does not move, you have paid for activity rather than customers.
Common questions
- Can community building replace paid advertising for a web3 project?
- No. Advertising buys first contact with people who have never heard of a project, at a price that can be raised or lowered immediately. A community converts and retains interest that already exists, and takes months to reach a useful size. Projects with no top of funnel find that an active Discord makes no difference to the numbers, because community does not create awareness at scale.
- Why do web3 projects rely on Discord and Telegram instead of ads?
- Largely because the major advertising platforms restrict or prohibit promotion of cryptoassets, exchanges, wallets and in some cases NFTs, with exemptions that require certification and vary by country. A project that cannot buy reach organises the reach available to it. The secondary reason is that holders have a stake in the project succeeding, which produces genuine word of mouth that ad audiences never do.
- What does it cost to run a web3 community?
- It is a headcount cost rather than a media cost. A public chat open around the clock needs moderation around the clock, because impersonation, phishing in direct messages and compromised webhooks are continuous rather than occasional. It also needs senior time for the questions that cannot be delegated, such as explaining a delay or a decision that lost people money.
- How can you tell if a crypto community is real or just an audience?
- Stop posting for three days. If members talk to each other, answer each other's questions and nobody assumes the project has died, it is a community. If the room goes silent, it is an announcement list. A second test is to read the last two hundred messages and count how many concern price, since a room that only discusses price cannot explain the product to anyone and empties during a drawdown.
- Are airdrops an effective marketing channel?
- They reliably produce wallets and rarely produce customers. Incentive campaigns attract farming, where participants complete the minimum action, claim, and leave when the reward ends, often across many wallets controlled by one person. They are a poor fix for an awareness problem. The diagnostic is simple: if wallet count rises while retention and repeat use do not move, the spend bought activity rather than demand.
