Dead Audience Detection: Five Checks Before You Spend

Run these five checks before you sign. They take ten minutes, require no software, and surface the kind of audience bloat that turns a $2,000 partnership into zero conversions. The work is diligence, not detective theater—you're protecting payback period, not litigating someone's character.

Most follower inflation happens gradually. A creator buys a batch early to clear a psychological threshold, or they use a growth service that mixes real follows with bot armies. The account continues to post, engagement drifts downward, and the numbers on the profile page become fiction. Your job is to spot the gap between the audience size and the audience that will see your product.

Check engagement shape across ten posts

Pull the last ten feed posts. For each, divide total engagements—likes, comments, shares—by follower count. You want a range, not an average.

Healthy accounts show variation. A workout video hits four percent one week, a life update hits one percent the next. That variance reflects an algorithm responding to content type and posting time. If every post lands between 0.8 and 1.1 percent with metronomic consistency, you're likely looking at the same small group of real followers engaging while the rest of the audience sits inert.

Also watch for cliff drops. A creator at 95K followers whose engagement rate suddenly fell from five percent to one percent six months ago either changed content direction or bought a block of follows that diluted the active base. The latter is more common. If the content style hasn't changed but the engagement collapsed, the follower count became decorative.

Read thirty comments for sentence structure

Open three recent posts. Read the first ten comments on each. Ignore the creator's replies.

Real comments have texture. They reference specifics from the video, ask follow-up questions, tag a friend with context. Bot comments are smooth and generic: "Love this energy," "Always inspired," "Keep going." They work anywhere, which is the tell.

You're also checking for language mismatch. If the creator's bio and captions are in English, the audience should skew anglophone. A comment section dominated by single-emoji responses or generic phrases in unrelated languages suggests bulk engagement services. An occasional non-English comment is normal. Thirty percent of the comment section in a language the creator doesn't use is not.

Don't dismiss an account over five bot comments. You're looking for proportion. If more than half the visible comments could be pasted under any post on any account, the engagement is decorative.

Map follower growth against content cadence

Scroll the feed. Count posts per month for the past six months. Then check a follower-tracking site—most are free and show growth curves without login.

Growth should correlate loosely with output. A creator posting five times a week can grow faster than one posting twice a week, assuming the content works. What you're hunting for is the spike: a vertical leap of 10K or 20K followers in a two-week window with no corresponding viral post, no feature, no collab that would explain it.

Sometimes a creator will have a breakout post and you'll see the spike justified in the archive. That's fine. But if the growth chart looks like a staircase—flat, flat, vertical jump, flat, vertical jump—and the content during those jumps isn't noticeably different, you're seeing purchased blocks.

Also flag accounts that grow in round numbers. A jump from 48.3K to 60K in three days is a buy. A jump from 48.3K to 53.7K over two weeks is plausible organic growth.

Compare follower count to average view count

Look at the view count on the last ten pieces of video content—Reels, YouTube Shorts, TikToks. Calculate the median.

On a healthy account, median views sit somewhere between ten and fifty percent of follower count, depending on platform and niche. A creator with 80K followers pulling 8K to 30K views per video is operating within normal range. A creator with 80K followers pulling 1.2K views has a dead file.

The platform matters. Instagram Reels often show lower reach than TikTok because the algorithm weights different signals. But the ratio should still make intuitive sense. If you need to squint to imagine how the follower count and the view count belong to the same account, they probably don't.

This check is especially useful for YouTube. Subscriber counts on YouTube are sticky—people rarely unsubscribe—so an account that grew fast two years ago and then changed content direction may have a large subscriber base that no longer watches. That's not fraud, but it has the same effect on your CAC. You're paying for a number that doesn't convert to reach.

Run the geography sniff test on tagged posts and comments

Click through to the profiles of ten people who commented or liked recent posts. Check their bios, their locations if listed, their own follower counts.

You're asking: do these profiles look like they belong to the creator's stated audience? If the creator is a New York–based running coach talking to American marathoners, you'd expect a mix of American-looking profiles, some with running content of their own, most with a few hundred to a few thousand followers.

If instead you see a cluster of accounts with no posts, no profile pictures, and usernames that look like keyboard slams—"j8374hda" or "sarah_94837453"—you've found the bot farm. If you see a bunch of accounts based in a single foreign country with no obvious connection to the niche, you've found a bulk engagement service.

This test has limits. Plenty of real people lurk without posting or use private accounts. But in aggregate, the sample should resemble a community. If it looks like a call center, trust that instinct.

When the checks say no

If an account fails two or more of these tests, walk. You're not arbitrating truth; you're allocating budget to channels that can move product. An account with 60K followers and 900 views per video cannot do that, regardless of why the gap exists.

If an account fails one test but passes the others, request performance terms. Offer to pay per conversion or per landing-page click instead of a flat fee for the post. A creator with a real audience will often agree, especially if they're confident in their reach. A creator with a purchased audience will decline.

Some will say these checks are harsh or that they punish small creators who experimented early and regret it. Maybe. But your job is not to subsidize someone's renovation project. It's to buy access to people who will see your product and consider buying it. If the audience isn't there, the deal doesn't work. Post a deal on FitBodega once you've found someone who passes the sniff test—it's free, and brands can see what's live.

The short version

• Engagement shape matters more than engagement rate. Look for natural variance across posts, not metronomic consistency.
• Read the comments for texture. Generic praise that could live anywhere suggests automated or purchased engagement.
• Growth spikes need explanations. Vertical jumps without corresponding viral content or collabs indicate bought followers.
• Follower count should predict view count. A 10x gap between the two means the audience is decorative.
• Check the humans behind the engagement. If the profiles don't resemble a real community, they aren't one.