Influencer databases find creators. They don't vet them.
By Paul Taylor · Written from inside live campaigns, not from a content calendar.
$35,000. 18 influencers, some of them above 300,000 followers. A single month of spend. Zero sales, and not even decent traffic.
That is how a marketer with 15 years in the job described a creator campaign in public last year. The comments under the sponsored posts gave the rest away: other creators from the same engagement pod, praising each other.
Nothing in that campaign looked wrong on the day it was booked. The follower counts were real numbers on real profiles. The engagement rates were high. The creators sat in the right category. The filters said yes.
An influencer database is good at one job: telling you a creator exists. The money goes on a different question, what a video with your product in it will do, and no database column answers it.
Why database filters pick the wrong creators
A database ranks what it can read from the outside: followers, a headline engagement rate, a category tag, past sponsors. Those are also the easiest numbers to dress up.
Followers can be bought. Sellers of likes and followers compete on price in the open, and buyers compare them the way you would compare hosting plans.
Engagement can be traded. A pod is a group of creators who like and comment on each other's posts on schedule. From the outside, it reads as a lively audience.
A single video that took off can make a channel look far bigger than its next upload will be.
Sort by those numbers and the creators who work the numbers rise to the top. The ones whose audience actually listens sit further down, with smaller figures and no reason to inflate them.
How to vet an influencer before you pay
When we build a shortlist for a brand, a database is where the search starts, not where it ends. Four checks decide whether a name goes to the brand.
Recent videos, not the best ones. A channel's latest uploads tell you what your video is likely to get. Its peak tells you what happened once.
Comments read by a person. Real viewers ask about the product, push back and mention their own setup. A pod leaves short praise from accounts that look a lot like the creator.
Audience location from the creator. We ask the creator where their viewers are instead of guessing from the language of the channel. A brand that sells in one market gets nothing from views in another.
A reply with a price and dates. A name goes to the brand once the creator has answered. Until then it is a guess with a profile picture.
Three questions to ask about any creator list
Whether the list comes from an agency, a platform or your own team, ask this about each name before you approve it. Is the views figure from recent videos or from the channel's best moment? Who confirmed where the audience lives, and how? Has the creator replied with a price and dates, or is this still a row from a database?
None of these needs a new tool. They need someone to ask them before the budget goes out, not after the report comes back.
A database saves you the search. It does not do the selection. Before you approve a name, ask to see what sits behind it.