Why I Built Authority Lab

I got tired of being told what works on LinkedIn by people who were guessing.

Every piece of advice sounded the same. Post at 9am. Use a hook. Tell a story. Add a CTA. It was all confident and none of it was sourced. When I asked "how do you know?", the honest answer was always some version of "it worked for me once." That is not knowledge. That is a coincidence with good marketing.

So I did the annoying thing. I decided to check.

The problem with "what works"

LinkedIn advice is treated like it is settled. But nobody had actually measured it at scale. The rules were built on a handful of anecdotes, generalised into laws, and passed around until they felt like fact. Meanwhile I would watch two posts with the identical "correct" structure get wildly different results, and nobody could tell me why.

I did not want more opinions. I wanted a big enough pile of real posts that the patterns had to show themselves.

What I actually built

The Authority Index started as a database. Tens of thousands of LinkedIn posts from the best B2B creators on the platform, each classified across more than 100 fields: intent, format, hook type, word count, cognitive load, CTA type, and how it performed relative to that creator's own baseline.

That last part matters more than anything. A post with 200 likes means nothing on its own. It means something when you know it did 3x that creator's median. Performance only makes sense relative to the person who posted it, so that is how we measure it.

The moment I knew it mattered

A few weeks in, I found out my own engagement scoring had been wrong for weeks. I had been ranking posts by comments, because I value comments, and quietly penalising visual posts that get saved and reposted instead. The day I fixed it, the average post score jumped from 1,833 to 3,106. Overnight.

I had baked my own bias into the product and not noticed. The data did not care what I valued. It just showed me I was wrong. That is the whole point of the thing. Not to confirm what I already believe, but to catch me when I am believing something false.

What Authority Lab is, and is not

It is not an AI that writes your posts for you. There are enough of those, and they all produce the same beige content because they are all trained on the same average.

It is a thinking tool. It tells you what the data says, shows you where your content sits against it, and leaves the writing to you, because the writing is the one part that should sound like a person. The rest of this blog is me reporting what is in the Index, including the parts that contradict what everyone told you.