Per SociaVault's 2025 audit of 1.2M creator accounts, 37.2% of influencer followers are estimated fake. That maps to roughly $4.6 billion of advertiser spend per year routed to bot audiences.

You don't need a forensic accounting team to avoid being part of that number. You need 60 seconds, three browser tabs, and the discipline to walk away from creators whose math doesn't pass a basic smell test.

Here's the walkthrough we run on every creator before we put a brand's money on them.

The 60-second audit, step by step

Step 1: Engagement rate vs. follower count (10 seconds)

Open their grid. Look at the last 12 non-pinned posts (excluding any obvious giveaway). Average the likes. Divide by follower count.

Healthy bands by platform in 2026:

• Instagram: 1.2%–4.5% for 100k–500k accounts. Below 0.8% is suspicious. Above 8% is also suspicious (engagement pods or bought engagement to mask bought followers).
• TikTok: 4%–12% on average view-to-like ratio. Below 2% is suspicious.
• YouTube: likes-to-views around 2.5%–5%. Below 1% suggests subscriber inflation.

A creator with 380k Instagram followers averaging 800 likes per post is not a creator. They are a wallet draining your budget into a void.

Step 2: Comment quality scan (15 seconds)

Open the top 3 posts. Scroll the comments. Look for:

• Generic emoji-only comments (three flame emojis, heart-eyes) at >40% of the comment count = bot trail
• Comments in languages that don't match the creator's stated audience at >25% volume = bought from a comment farm in another market
• Repeated phrases across multiple commenters ("amazing post! love it!" appearing verbatim three times) = automated engagement service
• Zero comment threads (no replies between commenters or from the creator) = the comments aren't from a real community

Real audiences argue with each other in the comments. Bots don't.

Step 3: Growth curve shape (15 seconds)

Use a free tool (Social Blade, HypeAuditor's free tier, ViralFindr) and pull their 12-month follower growth chart.

What you want to see: a gently sloping upward line with occasional bumps from viral posts. What you do not want to see:

• Vertical cliffs of +20k followers in 48 hours with no corresponding viral content
• Stair-step patterns where growth is flat for weeks then jumps by 5k repeatedly (bought in batches)
• A massive spike followed by a decay as the platform deletes the bot batch

Real growth is messy and roughly continuous. Bought growth has a signature.

Step 4: Audience demographic match (10 seconds)

If they share an audience country split (any reputable creator can produce one from their analytics), check it against where they actually post from and who they collaborate with. A creator based in Los Angeles, collaborating with US brands, who claims a 62% audience in Indonesia, has bought followers from a click farm in Jakarta. We've seen this exact pattern five times this year alone.

Step 5: Story view ratio (10 seconds)

Ask for a screenshot of the last 5 Story view counts. Real ratio of Story views to followers in 2026: 3%–10% depending on posting frequency. If the creator has 200k followers and 800 story views, the followers aren't watching them because most of the followers aren't real people.

A real audit walkthrough

Here's one we ran in March 2026 (creator anonymized; vertical: beauty; pitched to us at $9,200 per Reel):

• Followers: 412k Instagram
• Avg likes (last 12 posts): 1,640
• Engagement rate: 0.40% (red flag — well below the 1.2% floor)
• Comment scan: 67% emoji-only, 31% in Portuguese on an English-language US creator's account
• Growth curve: flat for 8 months, then +180k followers across two 36-hour windows in October 2024
• Stated audience: "78% US female 25–44"
• Actual audience pattern based on top-engaged followers (sampled 200): 71% Brazilian accounts, most with <30 followers themselves

We declined. They were re-pitched to us six weeks later by a different agent at a different agency, at $11,000. Same account.

What "good" looks like

A clean creator we approved the same week (vertical: beauty; 188k Instagram):

• Engagement rate: 3.2%
• Comment quality: real questions, real threads, creator replying
• Growth curve: organic upward slope, 4–6k per month, two clear viral bumps tied to specific posts
• Story views: 7.4% of followers consistently
• Audience: 64% US, 18% UK, 11% Canada — matches her collaborator graph

She charged $4,400 per Reel. Less than half the price of the fraud account. That is the second-order tax of bought followers: not just that they don't convert, but that the fraudulent accounts are typically priced higher because they need to justify the fake number.

The shortcut: build a kill-list

Run this 60-second audit on every creator a manager pitches you. Track the failures. After 60 creators you'll have a personal database of suspect agents and managers who consistently rep fraudulent accounts. Add them to a kill-list. Don't take their pitches.

This is not glamorous work. It is the difference between a $4.6B industry-wide waste line and the rest of your budget converting at the rate it should.

The honest take

Most brands lose 15–35% of an influencer budget to fake-follower fraud not because the fraud is sophisticated, but because nobody on the marketing team has 60 seconds and the willingness to say no. Build the audit into your contracting process. Make it a gate, not a vibe check.

Then go look at the last three creators you approved. We'll wait.