How to Measure Crypto Community Quality, Not Just Size

To measure crypto community quality, track behaviour rather than headcount: how many people contribute week after week, what share of questions are about the product rather than rewards, whether members answer each other, what people create without being paid, and what happens to activity when incentives pause. A small community that scores well on those signals is worth more than a large one that only wakes up for a snapshot. Member counts tell you how many people joined. Quality tells you how many would stay.

I use a 12-point scorecard for this, which I'll walk through below. It takes about two hours a month to run, and it works on your own community or on someone else's.

Why is community size such a poor signal in crypto?

In most industries, a community grows because people want to be in it. In crypto, people often join because joining was a task. A quest asks them to enter the Discord, follow the X account and join the Telegram, and once the points are logged, most never speak again. Member counts only go up, because leaving costs effort and staying costs nothing.

So a Discord with six figures of members can be quieter than a 2,000-person Telegram where builders argue about your roadmap every day. The size number hides that completely.

the biggest brands in future will have a loyal community as opposed to just buyers. Shilika Jain on X, 27 Dec 2022

A community of buyers is really a list of people waiting for a price event. A loyal community is people who'd still show up if the token didn't exist. The scorecard is designed to tell those two apart.

The 12-point community quality scorecard

Score each of six signals from 0 to 2, using the last 30 days of activity. Total out of 12.

Rough reading of the total:

• 0 to 4: You have an audience assembled by incentives. It may still be useful for distribution, but don't call it a community in investor updates or press.
• 5 to 8: There's a real core inside a larger farmed shell. Find those people and build around them.
• 9 to 12: You have a community. Protect it from campaigns that would flood it.

These bands are judgement, not research. Treat the score as a way to compare your community against itself over time, not against other projects.

How do you sample questions without a data team?

The organic question mix is the most revealing signal on the card, and it's the one teams skip because it sounds like work. It isn't much.

The method I use:

1. Pick a fixed window, say the last 14 days, across your main channels.
2. Pull 50 questions at random, not the most-reacted ones. Scroll to a few random points and take every question you find until you have 50.
3. Tag each one with a single label: product use, technical or integration, roadmap or governance, rewards and eligibility, token price or listing, or other.
4. Count the product, technical and roadmap tags together. That's your organic share.
5. Repeat monthly with the same labels, so you can see the direction.

What you're looking for is the ratio, and especially how it shifts. If a new campaign launches and rewards questions jump from a small minority to most of the sample, you've learned exactly who that campaign brought in.

community building is a gut-trenching task! Shilika Jain on X, 16 Jan 2023

Part of why it's draining is that the community manager reads every "wen token" message individually. Sampling turns that noise into a number you can act on, and it gives the community team evidence when they push back on a campaign that will flood the channels.

What does support load tell you about community quality?

Support tickets are an underrated signal, because nobody files a ticket for a product they don't use. The topic of the ticket tells you what people are actually doing.

A farmed community generates support load about the reward: "my wallet was flagged", "my points didn't update", "why am I not eligible". A real community generates load about the product: "the bridge took longer than expected", "the SDK throws this error", "how do I set this up for my team".

Both kinds of ticket need answering. But if most of your support time goes on reward disputes, that's not a community problem to solve with more moderators. It's a signal that your incentives are attracting people who don't use the product, and that's worth raising with whoever designs the campaigns.

One practical step: tag tickets by root cause for one month. The split is usually obvious within a week, and it's the kind of evidence that changes minds in a growth meeting more than any opinion does.

How do you test a community in a quiet week?

Quiet-week behaviour is the closest thing to a stress test you can run without spending anything. Find the quietest recent week, one with no campaign, no announcement and no price drama, and compare it to a campaign week on three counts:

• Unique posters, not messages, because a few heavy posters can hide a collapse.
• Questions about the product, using the same tagging as above.
• Member-to-member replies.

If all three hold up reasonably, people are there for something other than the reward. If unique posters fall away almost entirely, the community exists only when you're paying for it.

most look at having noise but the real moat lies in bringing substantial value on-chain, bringing in real audience and real users Shilika Jain on X, 25 Nov 2025

I wrote that about the opportunity in India, but it applies to any community metric. Noise is easy to generate. The quiet week shows you what's under it.

Can you score someone else's community?

Yes, and this is where the scorecard earns its keep outside your own channels. Before you pay a KOL, co-host with a partner project, or commit to an ecosystem's community program, you can run a lighter version on their audience.

• For KOLs: look at who replies to their posts. Are the replies from accounts with history in your category, or from the same ring of engagement accounts on every post? Do their followers ask real questions in the comments? My KOL vetting checklist covers the fraud side in detail.
• For partner projects: join their Discord or Telegram for a week before announcing a collaboration. Sample their questions the same way you'd sample your own.
• For ecosystem programs: ask what share of participants in the previous cohort are still active. If nobody knows, that's an answer too.

This matters most when money is involved. Founders I've spoken to have been asked for large token allocations by creators, and the question that rarely gets asked in those conversations is whether the creator's audience brings real users at all. When I build KOL campaigns, the audience quality check happens before any rate conversation.

When doesn't this scorecard apply?

A few situations where the card will mislead you:

• Very early communities. With 200 members, a handful of active people can swing every signal. Wait until there's enough activity to sample.
• Developer-first products. Builders often talk in GitHub issues and private channels, not in a public Discord. Score where the conversation actually happens.
• Communities built around culture. Meme and mascot-led communities can score low on "product questions" and still be loyal and durable. I've written about why degen marketing sometimes outlasts big incentive campaigns. For those, weight repeat contributors and unprompted creation more heavily.
• Right after a token event. Price drives behaviour for a few weeks. Don't score during a launch or a crash.

The other trade-off is political. A low score after a big campaign can feel like an attack on the growth team. Present it as a baseline you'll improve together, not a verdict. The point of measuring quality is to decide where to spend effort, not to prove a campaign failed. For the broader approach to growing a community that holds up, see how I work on crypto community growth, and for channel choice, community-led growth on Discord, Slack and Telegram.

FAQ

How often should we score our community?

Monthly is enough for most teams, with an extra check after any large campaign. Keep the same sampling method and labels each time, so you're comparing like with like.

Should we remove inactive members to make the numbers look better?

Pruning won't change the quality score, because the card measures behaviour, not headcount. It can help moderators and reduce spam, but don't present a smaller member count as a quality improvement on its own.

Who should run the scorecard?

Whoever owns the community, with a second person from product or growth checking the question tags. Having someone outside the community team review it keeps the scoring honest.

Can a community score well and still have airdrop farmers in it?

Almost certainly. Nearly every crypto community has some. The scorecard doesn't try to remove them; it tells you whether there's a real core worth building around underneath.

Want an outside read on how your community actually scores? Book a 30-minute call and we'll run the card on your channels together.