Blog

We argue that most of what the market calls GEO is unproven. Here are the five strongest published works that say the opposite, and the conditions that would change our…

Nobody can tell you whether GEO works in the abstract. You can measure it on your own site, in thirty days, with a control group and repeated measurements.

55 percent of marketers now have a GEO budget line. Not one practitioner surveyed disclosed an amount. Here are our prices, and the list of what we refuse to bill…

Minus 58 percent, minus 70 percent, or minus 10 percent? Each camp measures a different thing and uses the same word. How to read the numbers on traffic decline.

An audit of 712 real queries across four engines measures that roughly one cited source in six is generated content. GEO mechanically rewards what degrades the corpus.

Injection attacks work on isolated systems. In production, a testbed of 1,000 black-hat sites measures 99.78 percent neutralization. The real weakness is somewhere else entirely.

Generative engines systematically favor third-party sources over the content you publish on your own site. Your GEO budget is therefore aimed at the wrong place.

Across 100,000 answers and 100 brands, three visibility tiers appear from the very first query. A small player does not start at zero. It starts at 11 percent.

A model's performance on a piece of information follows a U-shaped curve depending on where it sits in the context. Part of generative visibility is a placement artifact.

Visibility in generative search is a distribution, not a point. A single measurement, which is exactly what your monthly report contains, demonstrates nothing.

761,495 citation pairs analyzed: close to a third misrepresent the source they point at. What that means when the misrepresented source is you.

The Reuters Institute measures zero growth in chatbot use for news across the United States, the United Kingdom, France and Germany, while 55 percent of marketers open a GEO budget…