How to spot fake Instagram followers

Influencer marketing takes a large share of French brands’ budgets, and part of that money still goes to audiences that don’t exist. In 2026, fake Instagram followers, bought engagement and comment pods are still practices every brand and agency needs to spot before committing a budget.

This article is for media buyers, partnership managers and talent managers who want a concrete framework to vet a creator before signing. You’ll find 6 red flags, a comparison of detection tools, an 8-point vetting method and the French legal framework.

The state of influencer fraud in 2026

How big the problem is

Reliable numbers are rare. The most widely cited estimate comes from a 2019 study by Cheq and the University of Baltimore: fake followers were costing advertisers around $1.3 billion a year worldwide, roughly 15% of influencer spending. The figure is dated and the market has changed a lot since, but it gives an order of magnitude.

To measure your own exposure, run the numbers on an assumption. If 10% of an annual budget of €100,000 went to partly fake audiences, that would be €10,000 spent for nothing. The real figure depends on your creators and your checks: that is what the method below aims to reduce.

Why fraud persists

Three main reasons:

• Follower count is still a dominant criterion for many media buyers. As long as advertisers pay for volume, creators have a reason to inflate it.
• Detection tools are not used systematically. Many collaborations are still signed without any check at all.
• The techniques keep evolving: quieter engagement pods, more convincing bots, slower artificial growth to stay under the radar.

The French legal context

Law no. 2023-451 of June 9, 2023 regulates commercial influence in France. It does not address buying followers, but several of its rules apply directly to brands:

• Transparency: commercial content must carry the label “Publicité” (advertising) or “Collaboration commerciale” (paid partnership). Leaving it out is punishable by two years in prison and a €300,000 fine, and counts as a deceptive commercial practice.
• Written contract: mandatory, or the contract is void, with specific terms (deliverables, fee, rights and obligations). Decree no. 2025-1137 of November 28, 2025 sets the threshold at €1,000 excluding VAT, for contracts signed since January 1, 2026. How the threshold is calculated for repeat collaborations is worth a legal opinion.
• Joint liability: the advertiser, its representative, the influencer and the influencer’s agent are jointly liable for damage caused to third parties while carrying out the contract.

For brands, that means two things: you may have to answer for damage caused by a collaboration, and the written contract becomes the right place to put your requirements on audience authenticity.

The 6 types of fraud to know

Before the detection signals, here are the techniques in use:

• Buying followers: bots or inactive accounts sold in bulk to inflate an audience
• Engagement pods: private groups of creators who trade likes and comments to fake natural activity
• Buying views: artificial views on Reels or TikTok to trigger the algorithm
• Buying comments: generic comments ordered to keep up appearances
• Follow/unfollow: following accounts in bulk to get follows back, then unfollowing
• Engagement rate manipulation: a mix of the techniques above to reach flattering engagement levels

The most skilled fraudsters combine several techniques to stay below the threshold of the simplest tools.

The 6 red flags of fake Instagram followers

Signal 1. An engagement-to-follower ratio that doesn’t add up

The first indicator, and the fastest to check. The ranges below are rough benchmarks to adapt to your industry and content format, not published standards:

Account size	Typical engagement	Look closely if
Under 10K followers	5 to 9%	< 3% or > 15%
10K to 50K	3 to 6%	< 2% or > 10%
50K to 200K	2 to 4%	< 1.5% or > 8%
200K to 1M	1 to 2.5%	< 0.8% or > 6%
1M and up	0.5 to 1.5%	< 0.3% or > 4%

• Under 10K 5 to 9%
• 10K to 50K 3 to 6%
• 50K to 200K 2 to 4%
• 200K to 1M 1 to 2.5%
• 1M and up 0.5 to 1.5%

Typical range Look closely Rough benchmarks, not published standards. Scale from 0 to 16%.

An unusually low rate can point to an inactive or bought audience. An unusually high rate can give away an engagement pod. Either way, take a closer look.

Signal 2. Comment quality

Fake comments are easy to recognize. Go through the creator’s last 20 posts:

Suspicious comments:

• Emojis only: 🔥🔥🔥 / ❤️❤️ / 😍😍
• Generic phrases unrelated to the content: “Nice!”, “Amazing!”, “Great post!”
• The same phrases repeated across several posts
• Comments in a language unrelated to the stated audience
• Commenter profiles with no photo, no posts, recently created

Genuine comments:

• Specific questions about the content
• References to other posts or to the community
• Conversations between followers
• The creator replying to comments (a good sign)

Signal 3. The follower growth curve

Natural growth follows a fairly steady curve, with spikes tied to a post that takes off or a press mention. Artificial growth leaves traces:

• Sharp spikes followed by a plateau or a drop: a one-off follower purchase
• Perfectly linear growth over weeks: automated follows
• Big drops followed by a rebound: Instagram purging fake accounts, then a new purchase

Tools that show growth history, like Social Blade or HypeAuditor, make these anomalies stand out.

Signal 4. Audience demographics that don’t match

A Paris-based creator who posts in French about local food should have a mostly French, French-speaking audience. If their stats show 60% of followers in countries unrelated to their content, that’s a strong signal.

Mismatches to watch for:

• Location: an audience concentrated in countries you wouldn’t expect for the content
• Age: an unrealistic split (everyone between 18 and 24 when the content targets 35 to 50 year olds)
• Gender: 95% women on a men’s lifestyle account, for example

This data requires access to the platform’s native stats: the creator shares them directly (screenshot, screen share) or through a tool connected to the platform via its official API.

Signal 5. The following-to-follower ratio

An account with 50,000 followers that itself follows 45,000 people should get your attention. Follow/unfollow means following in bulk to get follows back, then unfollowing. It inflates the count with no real engagement.

As a rule, a creator people follow for their content has far more followers than accounts they follow. Above 10K followers, a ratio of at least 10 to 1 is a common benchmark.

Signal 6. Past campaign performance

If you can see results from previous campaigns (some creators include them in their media kit), compare:

• The announced reach against the results you observed on your own metrics
• The click-through rate on tracked links
• The conversion rate on promo codes

A creator with 200K followers who drives 10 clicks on a story link has a weakly engaged audience. A creator with 30K followers who drives 800 will likely be more profitable, with almost 7 times fewer followers.

Detection tools compared, 2026

Tool	What it does	Data source	Price (October 2026)	Best for
Social Blade	Growth history	Public data	Free for basic stats	A quick first pass
HypeAuditor	Estimated audience audit	External analysis + estimates	Limited free version, paid beyond that	Pre-qualifying at volume
Modash	Audit + creator database	External analysis + estimates	From about $199/month	Agency research and audits
Kolsquare	Full platform	External analysis + estimates	On quote	Large campaigns, management
Reachy	The creator’s media kit	Official APIs, account connected by the creator	Free during the beta	Reading up-to-date stats from a creator who agrees to share them

Prices change often: check them on each vendor’s site.

Two different approaches. External audit tools (HypeAuditor, Modash, Kolsquare) analyze an account from the outside and estimate its audience quality without asking the creator for anything. Reachy starts from the other side: the creator connects their accounts, and their kit shows the numbers sent by the platforms’ APIs: followers and engagement, plus audience demographics on Instagram and YouTube. That rules out edited screenshots and numbers copied by hand. Reachy does not score how authentic an audience is, though: if a creator bought followers that Instagram hasn’t removed yet, they show up in the official numbers. And the bio, rates, formats and past brands are still entered by the creator.

A reasonable combination:

• Social Blade for the first pass
• HypeAuditor or Modash to spot audience anomalies
• A kit connected to the platforms, like Reachy’s, to read up-to-date stats before signing

The 8-point vetting method before any collaboration

Here is the checklist to follow, from the first pass to the signed contract.

1. Check the engagement-to-follower ratio

Calculate the average engagement rate over the last 12 posts. Compare it with the benchmarks in Signal 1. A clear gap, up or down, calls for a closer review.

2. Inspect comment quality

Open 5 random posts and read the first 20 comments on each. If more than 30% are generic (emojis only, “nice”, “love it”), take a closer look.

3. Analyze the growth curve

Look up the creator’s handle on Social Blade. Spot sharp spikes or growth that is too linear. Save a screenshot in the collaboration file.

4. Check that the demographics match

Ask the creator for screenshots of their native stats (Instagram Insights, TikTok Analytics) or a media kit connected to the platforms, a Reachy kit for example. Check that the demographics match your target.

5. Calculate the following-to-follower ratio

A quick check on the profile. If the creator follows more than one account for every five followers, that’s a red flag.

6. Ask for past campaign results

The creator should be able to share at least one previous campaign with its results (clicks, conversions, reach). A refusal or a vague answer is a red flag.

7. Run a qualifying video call

15 minutes with the creator, sharing their screen on their native stats live. It’s the simplest way to confirm or rule out the previous signals, and a screen shown live is hard to fake.

8. Add protective clauses

A written contract is mandatory above €1,000 excluding VAT. Use it to include at least three clauses:

• A statement from the creator on the authenticity of their audience
• A commitment not to buy followers, views or engagement
• A termination and refund clause if fraud is found during or after the campaign

Have your legal counsel review these clauses. This method adds about 30 minutes per collaboration to your buying process.

What fake followers cost your influencer budget

Take a fictional example to show how it works. A campaign with a creator who has 100,000 followers, a stated engagement rate of 5% and a fee of €3,000 for one sponsored post.

If the audience is entirely real:

• 100,000 followers who might see the post
• 5,000 interactions
• Cost per interaction: €0.60

If 30% of the audience is fake (the example’s assumption):

• 70,000 real followers
• 1,500 to 2,000 real interactions, once fabricated engagement is removed
• Cost per real interaction: €1.50 to €2

In this example, each real interaction costs you about 3 times more. Repeated across 30 creators in a year, the gap would add up to €18,000 to €25,000. Your own numbers will differ, but the order of magnitude shows what a check can save you.

Ask for a kit whose numbers come from the platforms

Auditing every creator yourself takes time. A simple step is to ask for a media kit whose stats come from the platforms, instead of a PDF or screenshots.

That’s what Reachy offers creators: they connect their Instagram, TikTok, YouTube, Facebook or Twitch accounts, and their kit, a public web page, updates itself from the official APIs. For you, as a brand or agency, that brings three things:

• Numbers nobody typed in by hand: followers, engagement and, depending on the network, demographics come from the platforms, not from a screenshot that can be edited
• Recent numbers: the shared link shows the latest known stats, not the ones from six months ago
• A faster first pass: reading the kit is often enough to rule out a profile that doesn’t fit your target

Keep in mind what this kit doesn’t do. It doesn’t detect fraud and doesn’t replace an audit: an official number can include fake followers, and rates, formats and past brands are declared by the creator. It complements the checks above.

How to build vetting into your media buying process

Three levels, depending on your volume.

Level 1: case-by-case checks

For small teams running fewer than 20 collaborations a year. Apply the 8-point checklist to every creator, with a simple tracking file. Time: about 30 minutes per collaboration.

Level 2: a formal process

For agencies managing 50 to 200 collaborations a year. A brief template that covers the 8 points, a short team training with regular reviews, and a request for a platform-connected kit from every creator. Time: a few weeks for the team to build the habit.

Level 3: dedicated tools and systematic requirements

For advertisers running several hundred collaborations a year. A subscription to a pre-qualification tool (HypeAuditor or Modash), platform-connected stats required above €1,000, and an independent audit for the largest budgets. Cost: built into the operating budget.

FAQ for brands and agencies

Conclusion: check before you sign

Influencer fraud won’t disappear in 2026. Brands that check before signing are hit by it far less. The method is known, the tools exist, and the written contract, now mandatory, gives you a framework to formalize your requirements.

Before every collaboration above €1,000, ask for stats that come from the platforms, apply the 8-point checklist and write your protective clauses into the contract. Half an hour up front can save you wasted spend down the line.

---

Read next: How to create a media kit: the complete beginner’s guide, The best tools for content creators in 2026, The most profitable niches for content creators in 2026. And to generate your media kit from your real stats: Reachy.