How to Spot Fake TikTok Engagement Before You Pay a Creator

A creator with 400,000 followers quotes you $3,000 for a post. Their videos look fine. Comments are positive. You have about ten minutes to decide whether that audience is real.

Follower count tells you almost nothing here. Buying 100,000 TikTok followers costs under $200, and engagement pods make the likes look plausible too. What does hold up is the relationship between the numbers, because that is much harder to fake convincingly across every metric at once.

Here are four checks that take minutes each, and how to run them across a whole shortlist instead of one profile at a time.

Check 1: engagement rate against the follower band

Engagement rate is the first filter, but only if you compare it to the right benchmark. Rates fall as accounts grow, so judging a 500,000 follower account against a 5,000 follower one produces nonsense.

Engagement rate = (likes + comments + shares) ÷ followers × 100

Rough bands on TikTok:

Bought followers push the denominator up while the numerator stays put, so the rate collapses. An account with 400,000 followers and a 0.6% engagement rate is the clearest signal you will get.

The opposite is also worth noticing. A rate far above the band, say 25% on a 200,000 follower account, usually means bought engagement rather than a phenomenon. Real virality shows up as one or two outlier videos, not as a uniformly excellent average.

Check 2: the comment to like ratio

This is the check most people skip, and it is the one that catches pods.

Comment ratio = comments ÷ likes × 100

Genuine TikTok audiences comment on roughly 0.5% to 2% of the likes they give. Below about 0.3%, something is off: likes are the cheapest thing to buy and the easiest to automate, so purchased engagement skews heavily toward them.

Above 5% is worth a look too. Unusually high comment ratios often mean an engagement pod, where a fixed group comments on each other's posts. Open the comments and read them. Pods produce generic text ("🔥🔥", "amazing!", "love this") from the same handful of accounts across multiple videos.

Check 3: consistency across recent posts

Pull the last 20 videos and look at the spread, not the average.

A real account is uneven. Some videos land, most do not, and the distribution is wide: a 50,000 view median with a 400,000 view outlier is completely normal.

Purchased engagement is suspiciously flat. If every video sits between 48,000 and 52,000 views, nobody is buying that pattern from an algorithm. They are buying it from a panel.

The cheap way to quantify it: divide the highest view count by the median. Real accounts frequently show 5x or more. A ratio under 1.5x across twenty videos means the numbers are being managed.

Check 4: does the audience match the offer

The last check is not a number. If you are paying for a skincare campaign and the creator's engagement comes from a general meme audience, the engagement rate can be perfect and the campaign will still do nothing.

Look at what the top three videos are actually about. If none of them relate to the category you are hiring for, the audience you are buying is not the audience you want.

Running this across a shortlist

Doing this by hand takes ten to fifteen minutes per creator. For a shortlist of forty, that is a full day, and it is the same four calculations every time.

Everything above comes from public profile data: follower count, and the like, comment, share and view counts on recent posts. That is exactly what the TikTok Data API returns in a single call, so you can compute all four checks in a loop.

const HOST = 'tiktok-data-pro.p.rapidapi.com';

async function vet(username) {
  const res = await fetch(
    `https://${HOST}/profile?username=${username}&limit=20`,
    {
      headers: {
        'x-rapidapi-key': process.env.RAPIDAPI_KEY,
        'x-rapidapi-host': HOST,
      },
    }
  );

  const { data, posts } = await res.json();
  if (!posts.length) return null;

  const sum = (key) => posts.reduce((t, p) => t + (p[key] || 0), 0);
  const avgLikes = sum('likes') / posts.length;
  const avgComments = sum('comments') / posts.length;
  const avgShares = sum('shares') / posts.length;

  const plays = posts.map((p) => p.plays || 0).sort((a, b) => a - b);
  const median = plays[Math.floor(plays.length / 2)] || 1;

  return {
    username,
    followers: data.followers,
    engagementRate: ((avgLikes + avgComments + avgShares) / data.followers) * 100,
    commentRatio: (avgComments / avgLikes) * 100,
    outlierRatio: Math.max(...plays) / median,
  };
}

Then flag anything that fails a threshold:

function verdict({ followers, engagementRate, commentRatio, outlierRatio }) {
  const floor =
    followers > 1_000_000 ? 1 :
    followers > 100_000   ? 1.5 :
    followers > 10_000    ? 2.5 : 3;

  const flags = [];
  if (engagementRate < floor) flags.push('engagement below band');
  if (engagementRate > 25)    flags.push('engagement implausibly high');
  if (commentRatio < 0.3)     flags.push('likes without comments');
  if (commentRatio > 5)       flags.push('possible engagement pod');
  if (outlierRatio < 1.5)     flags.push('views suspiciously uniform');

  return { flags, pass: flags.length === 0 };
}

Forty creators becomes a few seconds of runtime and a sorted list, and you spend your attention on the ones that flagged rather than on the arithmetic.

What this will not tell you

These checks catch bought followers, like farms and pods. They do not tell you whether an audience will actually buy your product, whether the creator is pleasant to work with, or whether they have posted something that will embarrass you later. Read the comments and watch a few videos before you sign anything.

They are also a point-in-time snapshot. An account that looks clean today may have bought followers last year and grown into them since. If a creator matters enough, track the ratios over a few weeks rather than judging on one pull.

Used properly, this is a filter rather than a verdict: it removes the obviously bad options quickly so your judgement goes where it counts.

---

Want to run these checks yourself? The TikTok Data API returns follower counts and per-video engagement for any public account, with 25 free calls and no credit card. There is also a free engagement rate calculator if you only need to check one creator.