How to Vet a Micro-Influencer Before You Hire Them: The 2026 Checklist

Before you pay a micro-influencer, you want five questions answered: is this a real, reachable person; does their audience actually live where they say; is the engagement earned or traded; is there anything in their history that will burn your brand; and have you written the answers into the contract. Most brands check none of this until something goes wrong. Background-screening firm Kroll vetted more than 10,000 influencers and found about one in four posed a significant reputational risk to the brands working with them. Here’s the checklist we run, in order, before any contract.

What you’re actually screening for

Three different things go wrong with a creator, and they need three different checks.

Bought followers are the oldest trick. An account pads its count from a follower farm, so the number looks impressive and the audience underneath is hollow. Traded engagement is subtler. Creators join “pods” on Telegram or WhatsApp where dozens of members agree to like and comment on each other’s posts within minutes of publishing, which tricks the algorithm into pushing the post wider. Because the interactions come from real accounts with real histories, basic bot filters wave them through. Synthetic personas are the newest problem: a fully AI-generated “creator”, face and voice included, that never existed as a person at all.

This is happening because trust is getting harder to earn and easier to fake, and the tools to fake it keep getting cheaper. The encouraging part: all three leave evidence, and you can find it in about fifteen minutes per creator. Vetting comes after you’ve built a shortlist, so if you’re still at that stage, start with finding the right micro-influencers in Australia and come back here before you contact anyone.

The pre-hire vetting checklist

Run these in order, and stop at the first hard fail. There’s no point auditing engagement on an account you can’t confirm belongs to a real person.

1. Reachability. Can you get the creator on a live video call? Real creators reply and show up. Hard fail: dodges a call after agreeing to work with you.
2. Audience geography. Pull their audience-location breakdown. Does it match the market you actually sell to? Hard fail: a Melbourne food creator whose audience is 60% outside Australia.
3. Follower authenticity. Read the follower-growth graph for unnatural jumps. Hard fail: a vertical spike of thousands in one month with no viral post to explain it.
4. Engagement authenticity. Compare comments, saves and shares against the like count. Hard fail: hundreds of comments, near-zero saves or shares.
5. Brand-safety history. Search their last 12 months for anything that contradicts your brand. Hard fail: undisclosed competitor deals, or content you would never want sitting beside your logo.
6. Contract. Put the analytics promise and a synthetic-content ban in writing. Hard fail: they won’t warrant their own numbers.

The next sections are how to actually run the three checks that catch the most.

How do you check if the followers are real?

Two free checks catch most fake audiences.

The growth graph. Open the creator’s account in Social Blade and look at the follower curve. Organic growth is lumpy but gradual. A bought audience shows up as a near-vertical cliff in a single week or month, often followed by a flat plateau or a slow bleed as the platform purges fake accounts. One spike with a matching viral video is fine. A spike with nothing behind it is a purchase.

The ratios. On a 1,000–10,000 follower account, you’re checking two numbers:

• Engagement rate of roughly 2–4%, measured as likes plus comments over followers, averaged across the last ten non-pinned posts. Under about 1.5% is weak for this tier. A suspiciously high rate sitting on top of generic comments is worse, because that’s the pod signature.
• A sane following count. An account following 7,000 people to reach 8,000 followers is running follow-for-follow, not building an audience.

For audience location and age, the free tiers won’t show you enough; the creator’s own analytics will, which is the next check. If you’re vetting at volume, a paid audit on HypeAuditor or Modash maps audience authenticity for you. For the no-cost route, here are the free tools that handle the first vetting pass.

How do you confirm the engagement is genuine?

The most useful thing you can do costs nothing: ask the creator to screen-share or send screenshots of their own platform analytics. Genuine creators do this without flinching, because it’s how they win work. The ones who stall, who are “just updating the media kit”, or who send a two-year-old PDF are usually hiding the gap between their public numbers and their real ones.

Here’s the message we send, more or less word for word:

Loved your recent content. Before we lock this in, could you send a screenshot of your last 30 days of Instagram Insights or TikTok Analytics: reach, audience location, age, and engagement on your last few posts? Just want to make sure we’re a fit for your audience.

What you’re reading in the reply:

• Audience location that matches the market you sell to.
• Reach in a believable ratio to followers. A 5,000-follower account showing 50,000 reach per post is either genuinely viral or inflated, and you want to know which.
• Engagement that builds over a day or two, not a vertical spike in the first hour followed by a flatline. That spike-then-flat shape is the fingerprint of a pod.
• Comments with actual content. Strings of fire emojis and “so good!” on every post are pod currency. Real followers reference earlier posts, ask questions, and argue with each other.

How can you tell if an influencer is AI-generated?

This is the check most brands don’t run yet, and the one that will matter most over the next year. The defence is straightforward, because a fully AI-generated persona fails at being present and accountable in real time. Run this before you contract anyone you haven’t actually met:

1. A live, unscripted video call. Ask them to turn the camera on and talk through a recent post. A synthetic persona can’t, and a real creator does it in ninety seconds.
2. Backend analytics, screen-shared live. Have them open their analytics on the call rather than send an image, so you’re confirming the account is operated by the person in front of you.
3. A reverse-image check on their profile photo and a few posts. Stock or AI-generated faces tend to surface somewhere else, or nowhere at all.
4. A look for the tells in their video: hands and fingers that morph between frames, lighting and shadows that don’t match the room, lip-sync that drifts out, and comment sections written in flawless, oddly generic prose.

Australians badly overrate their radar here. The Commonwealth Bank found 89% of us are confident we can spot an AI-generated scam, while only 42% actually can, which is worse than a coin toss. Don’t trust the gut. Run the call.

What to put in the contract

Vetting only holds if it survives into the contract. Three clauses do the work:

• An analytics warranty. The creator warrants that the follower and engagement figures they showed you are genuine and not artificially inflated. This turns a gut read into a promise you can enforce.
• A synthetic-content ban. No AI-generated likenesses, voices or audiences without your written sign-off, with the right to require provenance on submitted assets.
• A termination-for-cause right if any of that turns out to be false, or if they breach the disclosure rules.

Disclosure itself is its own subject: the #ad placement, and the AANA and ACCC obligations that now land on you as the commissioning brand and not only on the creator. We cover that side in full in the brand-safety playbook, and you shouldn’t sign anything before reading it.

The faster way to run all of this

Everything above is the manual version of a job that, done properly, takes fifteen to twenty minutes per creator. For one or two creators, do it by hand. For fifty, it’s a part-time job, and it’s the part most brands quietly skip right up until a campaign goes sideways.

That’s the reason we built these checks into Mega Donkey as a condition of entry, not a feature you switch on. Every creator clears follower, engagement-rate and bot-and-pod detection before a brand can see them, and a separate safety scan runs on anyone you shortlist. The check that earns its keep most, in our experience, is the audience-geography read paired with the live-analytics request: they’re the two things a fraudster can’t fake cheaply, which is exactly why they catch the most.

If you’d rather start from a pool of Australian creators who’ve already passed the fraud and engagement checks, that’s what verified creator discovery is built for. And if you’re still working out where to look, our guide to finding micro-influencers in Australia covers the find step before this vetting one.

Frequently asked questions

What should you check before hiring a micro-influencer?

Five things, in order: that the account is a real, reachable person (a live video call confirms it), that the audience is based where you sell, that the engagement is earned rather than traded in a pod, that their last year of content is brand-safe, and that the contract makes them warrant their numbers. Skip the contract step and the rest is unenforceable.

How do you ask a creator for their analytics without offending them?

Frame it as fit, not suspicion: ask for a screenshot of their last 30 days of reach, audience location and engagement “to make sure we’re right for your audience.” Genuine creators send it straight back, because sharing analytics is how they win work. Stalling, or a years-old media kit, is the red flag.

What engagement rate should a real micro-influencer have?

For a 1,000–10,000 follower account in Australia, roughly 2–4% (likes plus comments over followers) is healthy. Below about 1.5% is weak for the tier. A very high rate paired with one-word or emoji-only comments usually points to a pod, not a fanbase.

How do you spot an AI-generated or virtual influencer before you pay?

Insist on a live video call and a live screen-share of their platform analytics; a fully synthetic persona can’t do either cleanly. Then check submitted video for morphing hands, mismatched lighting and drifting lip-sync. Worth knowing: 89% of Australians think they can spot an AI scam and only 42% actually can, so run the call rather than trusting instinct.

Is it worth vetting creators yourself or using a platform?

Done properly, manual vetting runs 15–20 minutes per creator, which is fine for one or two and a part-time job for fifty. A verified marketplace runs the follower, engagement and bot checks before you ever see the creator, so you choose from a pre-cleared pool. For a handful, do it yourself; at volume, let the platform carry it.

Vetting a creator well is unglamorous and it’s the single highest-return twenty minutes in a campaign. Run the checklist, get the analytics on a live call, and put the promises in writing. Do that and the worst creators screen themselves out before they ever cost you a dollar. See how verified vetting works on Mega Donkey.