Follower count is the number everyone quotes and the least informative one available.
Followers are purchasable. Attention is not. The gap between the two is where most influencer budget disappears.
Start with the replies
The single most useful check takes about three minutes.
Open the creator’s last ten posts and read the replies. Not the count — the content.
A real audience argues. They ask follow-up questions, disagree, add context, reference earlier posts. A purchased or farmed audience posts emoji, generic praise, and the same four phrases repeatedly.
This is very hard to fake at scale and immediately obvious when you look. It tells you more than any analytics dashboard.
Engagement ratio, compared properly
Engagement rate in isolation means little, because it varies enormously by platform, niche and audience size.
The useful comparison is against similar accounts in the same category. A crypto-technical account with 20,000 followers has a characteristic engagement shape; one that deviates sharply from it is worth a closer look in either direction.
The pattern that should stop a booking: large follower count, tiny engagement. An account with 90,000 followers and twelve likes per post has bought something, and the fact that it is visible to you means it is visible to everyone evaluating your campaign afterwards.
Growth curves
Plot follower growth over time if the platform allows it.
Organic growth is lumpy but explicable — a post did well, they appeared on a podcast, a project they backed had a moment. You can usually point at the cause.
Purchased growth is a vertical line with nothing behind it. No post, no event, no explanation, and engagement that does not move with it.
Audience composition
Where is the audience actually located, and does it match?
A creator posting in English with an audience concentrated in markets where English is not spoken is a warning sign. So is an audience whose active hours do not correspond to the creator’s posting schedule.
For crypto specifically, check whether the audience is crypto-native at all. Plenty of general-interest creators take crypto sponsorships to audiences with no interest in the category, and reach into the wrong audience is indistinguishable from no reach at all.
Promotional history
Look at what they have promoted before, and what happened to it.
A creator with a history of promoting projects that collapsed is carrying that reputation, and it attaches to you. Their audience remembers.
Also worth checking: how often do they promote? A creator posting sponsored content weekly has an audience that has learned to discount it entirely.
Category fit over reach
The temptation is always to book the largest account available.
The creators who produce real results are usually mid-tier and specific — technical analysts, educators, developers with audiences, people running a visible portfolio. Their audience is smaller and it overlaps yours precisely.
Large general accounts deliver impressions. Specific ones deliver people who care.
Track everything separately
Every creator gets their own tracked link. Not a shared campaign code.
Without per-creator tracking you finish the campaign knowing you got signups and not knowing who sent the good ones — which means the renewal decision is guesswork, and you will renew the wrong people.
With it, sixty to ninety days later you can see which creators sent users who stayed. That group is usually a minority of the original wave, rarely the biggest names, and exactly who to build a longer relationship with.
Structuring the deal
Paid placement first, to gather the data. Performance-only arrangements at the start attract creators with the least confidence in their own audience, and the strong ones will not accept them from a product with no track record.
Then convert the creators whose referred users actually stayed onto affiliate or revenue-share terms. Those relationships compound, which a campaign never does.
Common questions
How do you vet a crypto influencer?
Check engagement ratio against follower count, look at who is actually replying, review their last twenty posts for whether the audience responds to substance or only to giveaways, and check their promotional history for projects that went badly. The strongest single signal is the quality of replies - a real audience argues with the creator, and a purchased one posts emoji.
What is a good engagement rate for a crypto KOL?
It varies enormously by platform and audience size, so the useful comparison is against similar accounts in the same niche rather than an absolute number. What matters more is the shape of the engagement: a 90,000-follower account with twelve likes per post has bought something, and a 4,000-follower account with sixty thoughtful replies has an audience worth reaching.
How do you spot fake influencer followers?
Sudden follower jumps with no corresponding content event, engagement that does not scale with follower growth, replies that are generic and emoji-heavy, follower accounts with no posts or profile pictures, and audience geography that does not match the creator language or market. Any one of these can be innocent. Several together are not.
Should you pay crypto influencers per post or on performance?
Start with paid placements to gather data, then move the creators whose referred users actually stayed onto an affiliate or revenue-share arrangement. Performance-only deals at the start attract the creators with the least confidence in their own audience, and the strongest KOLs will not take them from an unproven product.
Does Corum8 vet KOLs?
Yes, on audience data rather than follower count. We check engagement quality, audience composition, promotional history and category fit before anything is booked, put a tracked link behind every placement so each creator is reported separately, and manage the conversion into affiliate terms for those whose referred users stayed.
