What we’ve found, including the
parts that didn’t work.
Not a blog. Published method, published results, and a running account of what we’re learning by doing this on ourselves.
Editorial manifesto
Publishing first-hand material
is the entire point.
This is not a placeholder section to be filled later with generic posts. Publishing genuinely first-hand material is what makes a studio findable by the systems it sells visibility into — and it’s the only kind of writing a machine can’t already produce for itself.
Six pieces, in priority order. Each one has a job. Titles and standfirsts below are final; the articles themselves need writing.
Published pieces
The first
six.
“We had no AI visibility at all. Here’s the method, the baseline, and what happened.”
The one that matters most. We published the number before we changed anything, so the improvement — or the lack of it — would be checkable.
“How we measure AI visibility, in full.”
The complete method. How we build the question library, how many times we ask, how we report variation, and what we refuse to report at all. In a field where most practitioners say they don’t trust the available data, this is worth more than any claim about a result.
“Seven things AI visibility agencies claim that the evidence doesn’t support.”
The section from our homepage, expanded, with the reasoning shown. Written to be forwarded to somebody about to sign a contract.
“The data protection clock: what marketing teams have to do before enforcement lands.”
Timed to the deadline, written for the person who owns the risk register.
“Why your product pages are invisible and your homepage isn’t.”
The machine-readability gap, explained with page-type evidence. The most practically useful thing we can give away.
“What thousands of generated clips actually tell you about production cost.”
The distinctiveness argument. Written for anyone currently being sold speed.
We publish our methods,
including where they stop working.
If you want the underlying data behind this piece, ask us.