I have spent the last four years on both sides of a lot of brand deals. First as the person running Autosave, a car community in Singapore that grew from a Telegram group into something with 20,000-odd members. Then as the person building Influencees, which exists because of a problem I kept running into from the first side.
The problem is this. When a brand decides who to work with, the number they reach for first is follower count. It is the only number that is visible, standardised and free. And it is close to useless.
What follower count actually measures
Follower count measures accumulated reach. It tells you how many accounts, at some point, pressed a button. It does not tell you whether those accounts are still active, whether they see the content, whether they believe it, or whether they would act on a recommendation.
In a car community you notice this fast, because the feedback loop is short and physical. Someone posts a recommendation about a workshop, a tint shop, a tyre. If they are trusted, people go. If they are not, nothing happens, and everyone can see that nothing happened. After a while you learn who moves people and who just has an audience.
The two lists barely overlap.
The most valuable creators are often not the biggest
Some of the most valuable people I worked with at Autosave had a few thousand followers. They had spent years answering questions in comments, they owned the car they were talking about, and when they said something was worth buying, members bought it. Meanwhile plenty of larger accounts produced campaign posts that sank without trace.
This is not a novel observation — everyone in the industry says it. What nobody had was a way to act on it. If you cannot measure credibility, you fall back on the number you can measure, even knowing it is the wrong one. That is not stupidity, it is the absence of an alternative.
Why self-reported data does not fix it
The usual patch is a media kit: a creator sends over engagement rates, audience demographics, past campaign results. The problem is obvious once you have seen a few. The creator picks which numbers to show, which window to show them over, and which campaigns to mention. It is a CV, not an audit.
I do not think creators are being dishonest when they do this. Everybody presents their best case. But a market where every participant self-reports their own quality is a market with no price signal.
What we built instead
Influencees syncs performance data directly from Instagram and TikTok rather than accepting what a creator reports. That single change does most of the work: the same methodology applied to every creator, over the same window, without anyone choosing which slice to show.
On top of that sits the part I actually care about — credibility signals. Consistency over time. Whether engagement comes from an audience that behaves like a real audience. Whether the creator's stated niche matches what they actually post. None of this is exotic analytics. It is just what you would check manually if you had the time, applied to 1,200+ creators instead of five.
The goal is not to replace judgement. It is to make sure the judgement starts from a number that means something.
The uncomfortable part
A trust layer only works if it is willing to say unflattering things. A credibility score that never comes back low is a marketing asset, not a measurement. That tension is the whole business, and it is the part we will get judged on.
Communities taught me that trust is not a feature you ship. It is the accumulated residue of being right in public, repeatedly, including when it costs you something. I would like the infrastructure to reflect that.
