Skip to content
Sign in
Spot a fake

How to spot fake followers and views: 10 signals and a ratio

Updated · By Larpea

You spot fake followers and views by comparing three ratios: interactions to followers, interactions to views, and views to followers. Look for stepped growth, uniform views, generic comments and empty profiles. This page shows ten signals and a worked calculation. Ask for the platform's own report before paying for reach.

How to tell in 60 seconds. Divide interactions by views and by followers, then compare with the account's own history and the platform's report. The full method sits in how to spot a larp.

Why bought audiences exist, and why they matter

An audience number is currency: it earns brand deals, rankings and trust. That makes buying one tempting and checking one worthwhile. In the US, the FTC's rule on reviews and testimonials also covers fake indicators of social media influence at 16 CFR 465.8: as we read it, selling fake followers or views, or buying them to misrepresent influence for a commercial purpose, when the person knew or should have known they were fake, is an unfair or deceptive practice. Where the rule stops and what other countries do is in the US law guide, the UK guide and the Australian guide.

Ten signals

  1. Growth comes in steps. Why it matters: A real audience grows with content, and its curve is rough. Jumps of thousands overnight followed by a flat line look like a purchase.
  2. Views and interactions do not match. Why it matters: Many views with almost no likes, comments or saves means the viewers were not people looking at the video. Do the division in the worked example below.
  3. Views are suspiciously uniform. Why it matters: Ten posts all landing within a few percent of each other, whatever their content, are a pattern content does not produce.
  4. Comments are generic. Why it matters: Emoji-only or interchangeable comments, posted within seconds, from accounts that never post, are automated more often than human.
  5. Followers look empty. Why it matters: Sample a few accounts: default avatars, random-letter handles, no posts, thousands followed. A few are normal; a whole page is not. Profile pictures can be generated too (how to tell if a photo is AI).
  6. The audience does not match the account. Why it matters: A local business followed mostly from the other side of the world, or a creator in one language with an audience in another, is worth a question. Ask for the platform's audience report.
  7. Likes arrive without saves or shares. Why it matters: Real interest produces a spread of actions. A single action in bulk is bought.
  8. Spikes arrive at one hour with no source. Why it matters: A burst at the same time of day with no matching traffic from search, a share or another post suggests a service delivering on schedule.
  9. The analytics are a screenshot. Why it matters: An audience tab can be composed. Ask for a screen share where you see the page load, or the platform's own downloaded report.
  10. The pitch rests on a count. Why it matters: A brand deal or a sale justified by follower numbers alone ignores what matters, which is whether those followers buy. Ask for results of past campaigns from the brands.

A worked engagement ratio

The numbers below are invented for the example, in the style of a creator-analytics screen like the ones in creator analytics mockups, and they are not any real account's. The method is the point.

  1. Pick the posts. Take the last 10 posts, and leave out a one-off outlier so it does not flatter the average.
  2. Add the interactions. Count likes, comments, shares and saves on those posts. In our example, 3,600.
  3. Average per post. 3,600 divided by 10 is 360 interactions per post.
  4. Divide by followers. 360 divided by 48,000 followers is 0.75% engagement by followers.
  5. Divide by views. 3,600 interactions over 120,000 views is 3% engagement by views.
  6. Compare views to followers. 12,000 views per post against 48,000 followers is 25%. Then compare all three with the account's older posts, and with accounts of similar size.

There is no single healthy rate. What counts is the gap: an account whose three ratios are far out of line with its own history, or with similar accounts, is the one to question.

What to ask for instead

Ask for the platform's own audience and performance report, downloaded or shown live, for the last ninety days. Ask for results brands got from past work with the creator, with a contact you can use. Pay for outcomes (sales, sign-ups) and not for reach when you can.

If you are the one asked for proof

Share the platform's own report and expect to be asked for it. If you made a dashboard image for a video, label it as a mockup, and do not send it to a brand. Our guides to making a larp video and mock data for dashboards cover the honest use of these screens.

If you have been scammed

If you paid a creator or an agency because of inflated numbers, ask your payment provider for a dispute, keep the reports and the messages, and report it. Brands can also tell the platform about the account.

Official services only, as confirmed on the date shown. How much can be recovered depends on how fast you report.
CountryWhere to reportWhat it takesConfirmed on
United StatesReportFraud.ftc.gov (Federal Trade Commission)Scams and fraud of any kind.
United StatesIC3 (FBI Internet Crime Complaint Center)Cyber-enabled crime, including crypto theft.
United KingdomReport Fraud (City of London Police)England, Wales, Northern Ireland; Scotland: call 101.
CanadaCanadian Anti-Fraud Centre, online reporting systemOnline, anonymously if you wish; also tell local police.
AustraliaScamwatch (National Anti-Scam Centre, ACCC)Scams reported by the public; it does not recover money.
AustraliaReportCyber (Australian Signals Directorate)Cybercrime and hacked accounts.
IndiaNational Cyber Crime Reporting Portal and helpline 1930Call 1930 fast, with the transaction ID, then file online.

Questions

How can I tell if someone bought followers?
Look for stepped growth, followers with empty profiles, generic comments and a big gap between views and interactions. Then ask for the platform's own audience report, which is harder to compose than a screenshot.
What is a good engagement rate?
There is no single number. Compare an account's rate with its own history and with similar accounts. A sudden gap between views and interactions is more telling than one rate.
Can fake views be detected?
Often through patterns: views that are uniform, spikes at one hour and almost no interactions. A single signal proves nothing, but several together are a reason to ask for proof.
Is buying followers illegal?
In the US, the FTC rule at 16 CFR 465.8 covers buying or selling fake indicators of social media influence for commercial purposes. Other countries have their own rules; see the UK guide and the Australian guide.
Is a creator-analytics screenshot proof of reach?
No. Analytics screens can be composed; some apps make them for video props. The platform's own downloaded report or a live screen share is the evidence.

Keep reading