Most influencer vetting advice stops at fake followers, which is a bit like checking a used car for stolen plates and calling it a mechanical inspection. Fraud detection matters, and we will cover it properly in a moment. But the fake-follower problem is the easy one: it is measurable, tools catch most of it, and fifteen minutes of manual checking exposes the worst offenders. It is also not the mistake that quietly wastes most influencer budgets.
The expensive mistake is subtler. A creator can have a completely genuine audience that simply isn't your audience. Every follower real, every comment sincere, engagement healthy, and still not one person in that audience has a reason to buy what you sell. Nothing in a fraud report catches that, which is why good vetting runs three gates in order: is the audience real, is it yours, and is the creator safe to work with. Skip any one and you are gambling.
Engagement benchmarks below are typical ranges reported across industry sources in 2026 and vary by platform, niche and account size. Use them as reference points for comparison, not as pass-fail thresholds.
Gate one: is the audience real?
Fraud is common enough that this check is non-negotiable; industry surveys routinely find that a large majority of marketers have encountered some form of influencer fraud. Fortunately, most of it is visible without buying anything. Four checks catch the bulk of it.
Calculate engagement yourself. Do not take the rate from a media kit. Add average likes and comments across the last ten to fifteen posts, divide by follower count, and compare against the creator's tier. Smaller accounts naturally engage more, so a single universal benchmark is meaningless.
| Tier | Followers | Healthy range |
|---|---|---|
| Nano | 1K–10K | ~4–8% |
| Micro | 10K–50K | ~2.5–5% |
| Mid-tier | 50K–250K | ~1.8–3.5% |
| Macro | 250K–1M | ~1.2–2.5% |
| Mega | 1M+ | ~1–2% |
Read the result in both directions. Well below the range for that size suggests padded follower counts. Suspiciously high, such as double-digit engagement on a large account, can indicate purchased engagement or pod activity. TikTok generally runs higher than Instagram, and B2B niches lower, so judge against comparable creators.
Look at the growth curve. Steady, gradual growth is healthy. Vertical steps with no viral post to explain them are the classic signature of bought followers. Read the comments, properly. Twenty to fifty of them, not the top three. Real communities reference specifics from the post and talk to each other; purchased engagement produces interchangeable praise and emoji. Check audience geography. A creator whose content is in English about a local market, with a large share of followers in unrelated countries, is a familiar pattern.
Blinq
Digital business card platform for sharing contacts, capturing leads and syncing them to your CRM.
Best for: Networking, sharing contacts & capturing leads digitally
What you are actually buying
The reason authenticity matters commercially, rather than morally, is that it changes the price you are paying per real person reached. It is worth doing this arithmetic before you negotiate.
A creator quotes $2,000 for one post, with 200,000 followers and a media kit claiming 5% engagement.
→ Claimed engaged audience: 200,000 × 5% = 10,000 → implied $0.20 per engagement.
→ Your own calculation across the last 12 posts: 1.2% actual engagement.
→ Audience check: 28% of followers look suspicious → roughly 144,000 real people.
→ Real engaged audience: 144,000 × 1.2% ≈ 1,730 → about $1.16 per engagement.
Same fee, nearly six times the cost per real interaction. The creator may be worth $2,000 anyway, but now you are deciding that on real numbers rather than on the headline.
On the authenticity percentage itself, be realistic: every public account passively collects some bot followers. Tools generally treat under roughly 15% suspicious as normal background noise and flag above about 25% as worth investigating. What matters more than the pass-fail line is the absolute size of the real audience and whether the suspicious share is stable or climbing.
The real risk Fake followers are the easy problem. The expensive mistake is a completely genuine audience that simply isn't your audience: every follower real, and not one of them with a reason to buy what you sell.
Gate two: is it your audience?
This is where most campaigns are actually won or lost, and where most vetting stops short. A clean authenticity score tells you the audience is made of humans. It says nothing about whether those humans are your buyers.
Check audience demographics against your customer profile, using the creator's own analytics rather than their description of who follows them: age, location, gender split and language. A beauty creator with a genuinely engaged audience is worthless to you if 70% of that audience lives somewhere you do not ship. Check topical overlap too: does the creator's usual content sit naturally alongside your product, so a recommendation reads as continuous with what they normally talk about, rather than as a paid detour their audience will scroll past? The same fit logic drives creator-led social commerce, where the recommendation only converts if the audience already trusts the source.
The strongest check, if you can get it, is audience overlap, comparing their followers against your existing customers or against other creators you are considering. High overlap between two creators means you are paying twice to reach the same people. This is the same discipline as knowing your own first-party audience well enough to recognise a match. It is also why smaller creators so often outperform larger ones, a pattern we explored in why everyday creators outperform celebrity endorsements: a tightly matched audience converts better than a big, vague one.
Gate three: are they safe to work with?
The third gate is about risk rather than performance, and it is the one that produces the unpleasant surprises. Start with a content history audit: scroll back a year or two, not just the recent grid. You are looking for anything that would embarrass you by association, and for evidence that their values genuinely align with the way your brand speaks.
Then check commercial history. How much of their output is sponsored? A feed that is three-quarters ads has spent its credibility, and your post will land in a stream of them. Have they promoted a direct competitor recently, and is there anything in their existing deals that conflicts with yours? Disclosure practice is the other half of this: look at whether past sponsored posts were labelled clearly and early. A creator casual about disclosure is a compliance risk that attaches to your brand, not just theirs, and the enforcement mindset is the same one behind handling partner fraud and disputes, and the standards involved are the same ones we covered in FTC disclosure rules and where disclosures belong on the page.
Finally, judge professionalism from the first exchange: response times, whether they have a media kit, whether they ask sensible questions about your goals. The people who are organised in the enquiry stage are usually the ones who hit deadlines and brief requirements later, which matters far more in the long-term partnerships that now deliver most of the value.
Ask for the receipts
One habit upgrades everything above: verify claims against platform-native data. Ask for screenshots from the creator's own analytics, such as Instagram Insights, YouTube Studio or TikTok analytics, covering recent reach, audience demographics and engagement. These are far harder to fabricate than a media-kit PDF, and a creator who has nothing to hide will usually share them without fuss.
Cross-check what they show you against your own calculations. If the stated engagement rate does not match what you computed from public posts, ask why before you assume the worst, since reach can legitimately vary with format and platform changes, and a good creator will explain it. What you are testing is consistency and candour, the same due-diligence instinct behind vetting an affiliate offer before you promote it. On YouTube specifically, compare average view duration against engagement claims, since high engagement with very low watch time is a familiar mismatch, and the format differences matter as we noted in short clips versus long videos.
Manual checks or a tool?
For a handful of creators, do it manually. The full three-gate review takes roughly fifteen minutes per creator and costs nothing beyond attention, and free fake-follower checkers cover the basic estimate. Once you are assessing twenty or more creators a quarter, a dedicated vetting platform earns its keep by compressing that audit into a short review and adding audience-quality scores, demographic breakdowns and audience-overlap analysis that are painful to produce by hand.
Either way, keep the decision human. Tools produce a score; they cannot tell you whether a creator's voice fits your brand or whether their audience has a reason to care about your product. Treat the score as a filter that removes the obvious problems, then spend your judgement on fit. And once campaigns run, measure them against real outcomes rather than the reach numbers you started with, using the same scepticism we applied in attribution and an honest funnel audit.
The short version
Vetting influencers has three gates, and most brands only run the first. Gate one is authenticity: calculate engagement yourself across the last ten to fifteen posts, compare it against tier benchmarks rather than a universal number, check the growth curve for unexplained vertical steps, read twenty to fifty comments for genuine conversation, and look at audience geography. Treat under roughly 15% suspicious followers as normal and above about 25% as a warning, then redo the pricing maths on the real audience, because a fee that looks cheap against claimed reach often looks very different against actual engaged people. Gate two is fit, and it decides most campaigns: check audience demographics against your customer profile, topical overlap with your product, and audience overlap between creators so you are not paying twice for the same people. Gate three is safety: audit content history, sponsorship density, competitor conflicts and disclosure practice, and judge professionalism from the first exchange. Verify everything against platform-native screenshots rather than media-kit claims, use tools once volume demands it, and keep the fit judgement human.
Want creator partnerships that actually reach your buyers?
We help brands find, vet and run influencer programmes that deliver measurable results.
Explore Social Media Marketing →Frequently asked questions
How do you spot fake followers on an influencer account?
Start with the engagement rate, calculated yourself rather than taken from a media kit: add average likes and comments across the last ten to fifteen posts and divide by follower count. Compare that against the creator's tier rather than a universal number, since smaller accounts naturally engage more. Typical healthy ranges on Instagram run roughly 4 to 8% for nano accounts, 2.5 to 5% for micro, around 1.8 to 3.5% for mid-tier and 1 to 2.5% for macro and above, varying by platform and niche. Anything far below the range for that size suggests inflated followers, and anything far above it, such as double-digit engagement on a large account, can suggest purchased engagement. Then check three more things: the follower growth curve, where steady growth is healthy and sudden vertical steps with no viral post are a red flag; comment quality, reading twenty to fifty comments to see whether people reference specifics from the post or just post generic praise and emoji; and audience geography, where a large cluster of followers in a country unrelated to the creator's content is a common sign of purchased followers.
What percentage of fake followers is acceptable?
Zero is unrealistic, because every public account accumulates some bot followers passively without the creator doing anything wrong. Industry tools generally treat somewhere under about 15% suspicious accounts as normal background noise, and flag anything above roughly 25% as a problem worth investigating before you commit budget. The more useful way to read the number is in absolute terms rather than as a pass or fail: if a creator has 200,000 followers and 30% look suspicious, you are buying access to roughly 140,000 real people, and the price should reflect that. It is also worth checking whether the suspicious share is rising over time, which can indicate recent purchasing, versus stable, which suggests passive accumulation. And remember that a clean authenticity score proves only that the audience is made of real humans. It says nothing about whether those humans are your potential customers, which is the check that actually determines whether a campaign works.
What matters more than follower count when choosing an influencer?
Audience fit matters most, followed by engagement quality and content alignment. A creator with a smaller but tightly matched audience will almost always outperform a larger creator whose followers have no reason to buy what you sell, because reach you cannot convert is just an expensive impression. Fit means checking the audience demographics, location, and interests against your actual customer profile, not the creator's own description of who follows them. It also means looking at whether the creator's content topics genuinely overlap with your product's use case, so the recommendation makes sense to their audience rather than reading as a paid detour. After fit, look at engagement quality: are the comments real conversations that indicate trust, or generic praise? Then look at commercial history: how often they post sponsored content, whether past partnerships were disclosed properly, whether they have recently promoted a direct competitor, and whether their results for similar brands are verifiable. Follower count only tells you the size of the room, not whether the right people are in it.
Should you use influencer vetting tools or check manually?
Use manual checks when you are evaluating a handful of creators, and buy a tool when the volume makes manual work impractical. A careful manual review takes roughly fifteen minutes per creator and catches most obvious problems: you can calculate engagement yourself, scan the growth curve, read comments, and eyeball audience geography without paying for anything. Free checkers and analytics dashboards fill the gaps for basic fake-follower estimates. Once you are assessing twenty or more creators a quarter, a dedicated platform pays for itself by turning that multi-hour audit into a short review, and it gives you audience-quality scores, demographic breakdowns and audience-overlap analysis that are hard to produce by hand. Whichever route you take, one rule holds: verify claims against platform-native data. Ask creators for screenshots from their own analytics, such as Instagram Insights, YouTube Studio or TikTok analytics, because those are far harder to fabricate than a media kit, and cross-check the numbers they state against your own calculations.
