What six months of measuring my own AEO visibility actually showed
I tracked my own AEO visibility for six months. FAQ schema and verbatim question matching didn't move it, and the reason isn't the content.
Ben Rees - 3 November 2026
Standard AEO advice says to answer the question directly, use the buyer's exact wording and mark it up with FAQ schema. I've done all three on my own site and measured the result every few days since April. None of it moved my visibility.
What I measured
Since 19 April I've run a fixed set of queries through Google search and Google AI Mode, and, when I'm at the keyboard to run them, ChatGPT, Gemini, Claude and Copilot. Each run logs whether bjrees.com appears, in what form, and keeps a screenshot. The query list has been locked since August, so the trend measures the site, not my edits to the questions.
Five of those queries are the textbook definitional ones: "what is answer engine optimization", "what is generative engine optimization in B2B marketing" and three more like them.
As of this morning: 20 appearances in 1,035 readings. 1.9%.
On 20 October it was 20 out of 895. The denominator keeps growing. The numerator doesn't.
I did what the advice says, literally
HubSpot's AEO guide is a fair summary of the consensus: direct answers, structured data, conversational phrasing that mirrors the question.
On two posts I took that to the letter. My post on how to build a B2B marketing team carries FAQ schema whose questions are, word for word, two queries I track: "What does a VP of marketing do in a B2B company?" and "What are the key qualities of a great CMO?" Both sit at 0.6%.
My buyer-group post matches two of my three technical-buyer queries verbatim, in the schema and in the body copy. That category sits at 1.4%.
The tactic was implemented exactly. It did nothing I can measure.
It isn't the wording, and it isn't missing content
The comfortable explanations are that I've phrased things wrongly or haven't written enough. I checked both.
If my AEO posts were simply framed wrongly, real searchers would still find them under other phrasings. Search Console says otherwise. Over 90 days, three of the six posts I've written on AEO had zero impressions. The other three collected 19 between them, and 14 of those were branded site: lookups.
Then I asked the engines directly. This week I ran both definitional queries by hand. ChatGPT cited nothing at all for the first, and a scatter of loosely related pages for the second. Perplexity cited HubSpot, Salesforce, Meltwater, Forrester, Sitecore, IONOS and Coursera for "what is AEO". For the B2B version, it cited a list of agencies that sell GEO as a service.
bjrees.com appeared in none of roughly 24 source slots.
The same post proves both halves of this
The definitional answer is owned: platform vendors with content budgets, analysts, SEO tool companies, and now a crop of agencies whose business model is ranking for exactly these phrases. A one-person site does not out-corroborate that with better markup. That's an authority ceiling, not a formatting problem.
The interesting part is the build-a-team post. Its FAQ questions, the generic ones any HR blog could answer, sit at 0.6%. Its title query, "how to build a B2B marketing team", went from 0 appearances in 191 readings before I published it to 108 in 155 afterwards.
Same page, same schema, same domain. The difference is that the title query is answered from having actually built teams at Redgate and elsewhere, and the FAQ questions are answered from general knowledge. The pattern holds across the site: my scaling queries, "challenges of scaling B2B marketing past Series B" and the SMB-to-enterprise variants, sit between 36% and 82%. None of the companies in that Perplexity list can write those posts.
The limits
This is one site, one person and a query set I chose. A little over a thousand readings is enough to say AEO tactics don't quickly move a small site's definitional visibility. It isn't enough to say they never work for anyone.
My own instrument has also been wrong before. Until October my dashboard put AEO at 14%, because a labelling bug had filed three scaling queries under AEO. It's fixed, but it's a good reason to treat any single visibility number as provisional, mine included.
What I'm doing next
I've stopped writing definitional AEO content. The effort goes into what only I can write.
And one small test, built into this post. It contains figures that appear nowhere else: 1,035 readings, 20 appearances, 1.9%. Many retrieval pipelines still use term-weighting schemes like BM25 (Robertson and Zaragoza, 2009), which weight rare terms heavily. If an AI answer cites these exact numbers in the next 30 to 60 days, that's retrieval finding a study rather than a definition. I'll report what happens either way.
Related reading
People - the True Heart of Decent Marketing
My talk at The Marketing Meetup's Cambridge chapter argued that AI visibility is a brand problem, not a content problem, backed by 5,095 real citation checks showing two published articles that never moved a single query they were written for.
Testing whether I should reuse old content
I fit a hierarchical Bayesian model to test whether retitling old content for AI citation works, and a robustness check collapsed the one clean result it found.
My AI visibility went from zero to 26% in six weeks. The one category still at zero is the one this post is about.
Six weeks after publishing a zero-appearance rate across every AI platform, most categories moved. The one still at zero is AI Search Visibility itself.
I am invisible on every AI platform. Here is the data.
For ten weeks I ran an AI visibility measurement system across ChatGPT, Gemini, Copilot, and Perplexity. My appearance rate for non-branded queries is zero. Here is what the data actually shows.