The Five Checks That Stop You Paying for Fake Influencer Reach
Most influencer budgets are lost before the campaign starts, on the day someone approves a creator because the follower number looked good. Reach is the easiest metric to buy and the easiest one to present in a deck. These five checks take about ten minutes per creator and remove the guesswork.
Why does follower count tell you almost nothing?
Followers measure history. They tell you how many accounts pressed a button at some point, not how many people see a post today. An account can hold a large audience that stopped watching two years ago, and the platform will still show the number in bold at the top of the profile.
The number you actually pay for is attention on the next post. That is why a creator with 12,000 followers and a live audience often outperforms one with 200,000 followers who bought reach during a growth push. Judge the post, not the profile.
What does a real engagement rate calculation look like?
Take the last twenty to thirty posts, ignore the pinned ones, and work out engagement per post against follower count. Then look at the shape of the results rather than the average, because the average is where inflated accounts hide. Doing this by hand across a shortlist of forty creators is where most teams give up, and it is the one part of engagement rate vetting worth handing to a system.
Mean versus median, and why the gap matters
If the mean engagement rate is far above the median, one or two posts carried everything. That usually means a viral moment, a paid boost, or a burst of bought engagement. A healthy creator is boring in the best way: the numbers sit in a tight band post after post. Ask for the band, not the best week.
The comment quality read
Open the comments on five recent posts and read them. Real audiences ask questions, argue, tag friends and reference earlier posts. Bought engagement produces short generic praise and emoji strings with no thread underneath. This check costs three minutes and catches more than any score does.
How do you spot bought followers without a tool?
Automated fake follower flagging looks for the same signals you can check yourself, patterns a real audience does not produce:
- A sharp follower jump in one week with no post that explains it
- Follower growth rising while likes and comments stay flat or fall
- An audience located far from the market the creator posts about
- Comments arriving in a tight burst right after posting, then nothing
- Story views that sit far below what the follower count would suggest
Any one of these can have an innocent reason. Three together is a pass.
What should you check after the post goes live?
Vetting does not end at the contract. The gap between what was agreed and what appeared is where campaigns quietly lose value.
- Confirm the post went live on the agreed date and stayed up for the agreed window.
- Run deliverable verification against the brief: format, link placement, disclosure and tagging all have to match what was contracted.
- Compare the sponsored post against that creator’s own recent baseline, not against other creators.
- Record reach, saves and clicks while the campaign runs, because good influencer campaign reporting is impossible to rebuild three weeks later from screenshots.
- Note who delivered late. That list is worth more than any discovery database next quarter.
Frequently asked questions
What is a good engagement rate?
It depends on size and category, so a fixed number is misleading. Smaller accounts almost always show higher rates than large ones, and beauty behaves differently from B2B. Set your floor by benchmarking ten creators in your own category and size band, then measure everyone against that.
Should I trust the media kit a creator sends?
Treat it as a starting point. Media kits show the best month, which is fair, and they rarely show the median. Ask for a screen recording of the account insights covering the last ninety days instead. Creators with real numbers hand this over without friction.
Are micro creators always the better deal?
Not always, but they are usually the better test. Their audiences are tighter and their rates let you run six creators instead of one, which gives you data rather than a single result. Scale up behind whoever actually converted.
Can this whole process be automated?
The mechanical parts can. Pulling post history, computing real rates, flagging odd growth patterns and checking whether contracted posts went live are all repeatable, which is exactly the work an AI marketing teammate takes off a human. The judgement call on whether a creator fits the brand still belongs to you.
The short version
Pay for attention, not for audience size. Read the comments, check the median instead of the average, look at growth history for jumps that no post explains, and verify delivery after the campaign rather than trusting the invoice. Ten minutes per creator saves the budget that the deck never gets to explain.