What a Publisher Actually Controls
What SEO, AEO and GEO mean: search, answer and generative engines. They describe someone else's system; a publisher controls a short, checkable list.
Three acronyms turn up in the same conversation. Search engine optimisation, answer engine optimisation, generative engine optimisation. They are presented as three disciplines, or three layers, or three stages of a pipeline.
They have something more useful in common. Each of them names the behaviour of a system that somebody else operates, whose ranking logic is not published, whose retrieval behaviour changes without notice, and whose output you can observe only by sampling it.
What you hold
A publisher controls whether a page can be reached, what it declares itself to be, and how it is structured. It does not control whether any system retrieves, ranks or cites it.
The list of things you do hold is short:
- Access. Whether a crawler can fetch the page at all, and what your robots policy permits.
- Identity. Which URL is canonical, so the same content does not present as several competing pages.
- Listing. Whether the page appears in your sitemap, and whether that sitemap reflects what actually exists.
- Status. What codes your URLs return. A page that answers 200 when it should answer 404, or a redirect chain three hops long, is a fact about your site rather than a theory about anyone's engine.
- Linking. Whether your own pages point readers and crawlers toward the pages you consider important. It is one discoverability condition you directly control.
- Structure. Whether the page is legible to something reading it mechanically: headings that describe sections, a title and description that match the content, markup that says what the page is.
That is close to the whole list, and none of it is a growth tactic. It is whether the thing you published is fetchable, identifiable, listed, honest about its status, and readable.
- Checkable directly
What a publisher controls
- Access: whether a crawler can fetch the page
- Identity: which URL is canonical
- Listing: whether the page appears in your sitemap
- Status: what codes your URLs return
- Linking: whether your own pages point to it
- Structure: whether the page is legible to something reading it mechanically
What it does not control
- Whether any system retrieves it
- Whether any system ranks it
- Whether any system cites it
- Checkable directly
A publisher controls whether a page can be reached, what it declares itself to be, and how it is structured.
Why the distinction earns its place
The controls a publisher actually holds can each be checked directly, which is what separates them from the parts of discovery that can only be inferred.
You can fetch your own robots file. You can request a URL and read the status code. You can compare your sitemap against the pages that exist. You can view the page as markup. Each of those is a question with an answer today.
The rest of the subject does not work that way. Whether a particular system retrieves a particular page, how it segments the text, what it treats as authoritative, whether it cites you: those are properties of software you cannot inspect, and advice about them is inference presented as instruction. Some of that advice may be right. None of it is checkable by you, which is a different kind of claim and should be labelled as one.
How retrieval actually behaves once material reaches a system is its own subject, and it is not a publishing control.
What this changes
Before optimising for an engine, check the six things above on your own site. They take an afternoon, and they are worth checking rather than assumed, because nothing fails loudly when a canonical tag points at the wrong URL or a sitemap lists pages that were deleted last year.
What that gets you is not a ranking. It is the removal of the reasons a system might fail to see the page at all, which is the only part of this you can actually do.
Where this helps, and where it stops
- What it is
- A note on the publishing controls a site owner can verify directly, and the limits of what they establish.
- What it does not guarantee
- An account of how retrieval, ranking, embeddings, chunking, grounding or context assembly work, or a strategy for any particular engine.
- When the distinction matters
- Someone proposes optimising for a system whose behaviour nobody in the room can observe.