What percentage of influencers have fake followers?

Public estimates vary widely because most rest on small samples. Measured on our own index: of the 89,000+ creator profiles Koinon Link has indexed, 7% show bought-follower patterns strong enough that we quarantine them from search results (as of 2026-09). On TikTok specifically the figure is 12% of 36,000+ indexed profiles. This is our screening outcome — follower-growth shape, engagement ratio and view-to-follower ratio — not a platform-official statistic, and it counts accounts flagged, not accounts proven.

How do you calculate engagement rate?

The common version is (likes + comments + shares + saves) on a recent post divided by follower count, averaged over the last 5–15 posts. A per-view version — engagement divided by median views — is even harder to fake because views and interactions have to move together. The exact formula matters less than applying the same one consistently and comparing against the platform and follower-tier norm.

What is a "normal" engagement rate?

It varies by platform and size, and inversely with follower count: smaller accounts almost always engage higher. As rough public benchmarks, TikTok tends to run highest, short-form and Reels mid-range, and large Instagram/X accounts low single digits or less. Do not anchor on one universal number — anchor on the norm for that platform at that follower tier, and treat a big account sitting far below it as a red flag.

Why is a low engagement rate on a big account a red flag?

Followers can be bought in bulk; genuine likes, comments and saves cannot be manufactured at the same scale without looking wrong. So when a 500k-follower account gets a few hundred interactions, the most likely explanation is that a large share of the followers are bought or inactive. The ratio, not the raw follower number, is what predicts whether real people will see and act on a post.

Can engagement itself be faked?

Some of it — engagement pods and bot comments exist — but it is far more expensive and detectable than buying followers, and it usually leaves fingerprints: generic emoji comments, the same handful of accounts on every post, likes with no corresponding views. Cross-checking engagement against view counts and comment quality catches most of it. The only fully reliable confirmation is a tracked pilot measuring real actions.

Engagement Rate: The Metric That Exposes Fake Influencer Reach

Updated 2026-07-16

A follower count is the easiest number in marketing to fake — you can buy 100k of them for the price of lunch. Engagement is the hard part, because genuine likes, comments and saves come from real people who actually saw the post. That asymmetry is exactly why engagement rate, not follower count, is the number that tells you whether reach is real. Here is how to read it.

How common is this, measured? Of the 89,000+ creator profiles Koinon Link has indexed, 7% carry bought-follower patterns strong enough to be quarantined from search results (as of 2026-09). On TikTok the share is 12% across 36,000+ indexed profiles. Among the profiles that pass validation, median follower count is 5K+ and 7% publish an email or IM handle. These are our screening outcomes over our own index, not platform-official figures.

How to calculate it

The standard formula is interactions on a recent post — likes + comments + shares + saves — divided by follower count, averaged across the last 5–15 posts to smooth out one viral outlier. A stronger variant divides engagement by median views instead of followers: views and interactions have to move together, so faking this convincingly is much harder. Pick one formula and apply it consistently; the comparison is what matters, not the decimal.

What "normal" looks like

Two rules cover most cases:

Engagement falls as follower count rises. A 5k-follower KOC almost always engages at a higher rate than a 5M-follower celebrity. This is why mid-tier creators so often out-convert mega accounts — their audience is smaller but far more real and attentive.

The norm is platform-specific. TikTok tends to run highest, short-form and Reels mid-range, large X and Instagram accounts low single digits. Benchmark against the same platform and tier, never against a single universal figure.

The signal you act on is not the absolute number — it is the gap between a creator’s rate and the norm for their platform and size.

Why the ratio beats the follower count

Followers scale cheaply; real interaction does not. So the tell-tale pattern is a large account with interaction far below its tier norm: a 500k-follower profile pulling a few hundred likes almost certainly carries a heavy load of bought or dormant followers. In high-CPA verticals — iGaming, crypto, finance — that gap is not cosmetic, it is your budget: you pay for a number and reach almost no real people.

Where the ratio can still mislead

Engagement can be inflated too, via pods and bot comments — but it is expensive and leaves fingerprints: generic emoji praise, the same twenty accounts on every post, likes with no matching views. Cross-check engagement against view counts and read a few comment threads for real, product-specific talk. The only certain confirmation is a small tracked pilot measuring real actions — clicks that become signups, deposits or orders.

koinonlink.com builds this screen into discovery: it scores accounts on follower-growth shape, engagement ratio and view-to-follower ratio, and on Douyin it reads each creator’s 30-day interaction rate directly to catch high-follower / low-engagement accounts. Suspected bought-follower profiles are flagged and quarantined, so your shortlist starts from creators whose reach is real. It narrows the field; a tracked pilot confirms the winner.

This data is built for AI

No key, no account — one command to retrieve it. Creators can claim and correct their own page.

Open Koinon Link →