Content Structured for Citation Lesson 13 of 27

Length, depth, and when to stop

What you'll learn
  • Why "longer content ranks better" doesn't translate well to GEO
  • How to think about content length when you're writing for both AI and humans
  • A simple test for knowing when a page is the right length

The most common SEO myth, applied to GEO

A lot of SEO advice over the last decade has converged on a single recommendation: write longer. Long-form content ranks better. Comprehensive content ranks better. Two thousand words beats five hundred. Three thousand beats two thousand.

There’s a kernel of truth in this. Pages that cover a topic thoroughly often do better than pages that don’t. But the word-count framing is the wrong way to express it — and it falls apart almost completely in a GEO context.

AI systems don’t reward length. They reward clarity. A page that answers a question well in 400 words can be cited far more often than a page that answers the same question across 4,000. The longer page might do better in traditional search rankings; the shorter one might do better in AI citation. Neither is universally right.

The honest framing is this: write to the length the topic actually requires. Not more. Not less.

Why length matters less in GEO

A few things change when AI systems become a significant part of how people find your content.

AI lifts the answer, not the page. When ChatGPT cites you, it doesn’t lift your 3,000-word blog post. It lifts the two-sentence answer buried somewhere inside it. The other 2,950 words helped you rank in traditional search — they didn’t help you get cited.

Padding hurts citability. Long content tends to accumulate filler. Hedging phrases, throat-clearing introductions, sections written to hit a word count rather than to say something. AI systems skip filler more reliably than human readers do. A long page padded with weak content is genuinely worse for citation than a shorter, tighter version.

Burying the answer costs you. If the citable sentence is in paragraph 17, an AI may never find it. If it’s in paragraph two, it’s much more likely to be lifted. Long pages tend to delay the answer. Shorter pages tend to lead with it.

Comprehensiveness still matters — but it doesn’t mean length. A page that covers a topic comprehensively in 800 words is more useful than a page that pads the same coverage to 2,500. Comprehensiveness is about whether you’ve answered the real questions. It’s not about the word count beneath the answers.

What this adds up to: the SEO instinct to keep writing until you’ve hit a target word count is the wrong instinct for GEO. The right instinct is to keep writing until you’ve actually answered the question, then stop.

How to think about length, by content type

Different kinds of content have different natural lengths. A few rough guides.

FAQ answers. 50–150 words is usually right. Lead with the direct answer; expand only as much as the question needs. If an answer routinely runs longer than 150 words, the question is probably actually two questions and should be split.

Service pages. Whatever it takes to answer the buying questions clearly. For most service pages, that’s somewhere between 800 and 2,000 words. Longer than that usually means the page is trying to do the job of several pages.

Informational blog posts. Highly variable, but the strongest performers tend to land between 800 and 1,800 words. Long enough to cover the topic thoroughly, short enough that the citable parts aren’t buried.

Definitional and reference content. Often shorter than people expect. A page that exists to define one concept well doesn’t need 2,000 words. 400–700 is often plenty. The on-page SEO course’s individual lessons are a good model — focused, complete, finite.

Hero sections, summaries, and landing pages. Shorter is almost always better. The job is to make the reader move, not to inform them comprehensively.

None of these are rules. They’re starting points for what “the length the topic requires” tends to look like in practice.

A useful test

If you’re not sure whether a page is the right length, try this:

Read it through once. After each paragraph, ask yourself a single question: “Did that paragraph teach the reader something, or did it just take up space?”

If you find three or more paragraphs that didn’t teach anything, the page is too long. Cut them.

If you find that the page seems to be promising more than it delivers — the headings suggest depth that the body doesn’t actually provide — the page is too short. Expand the sections that are thin.

If most paragraphs teach something and the headings are honest about the depth, the page is approximately the right length, regardless of the word count.

This test ignores the conventional SEO framing entirely. It doesn’t care whether the page is 600 or 2,000 words. It cares whether the words on the page are doing work.

A small caveat about traditional SEO

If your content needs to do both jobs — rank in traditional search and be cited by AI — you’ll occasionally face a tension. A page might rank better if you padded it. The same padding hurts citation.

The honest answer is to err towards the shorter, tighter version. Search engines are getting better at recognising padded content; the long-tail SEO benefit of pure word count is shrinking every year. And the citation benefit of tight, well-structured content is growing. The two trends point in the same direction, even if they haven’t fully arrived yet.

If you have to choose, choose to be cited. The pages that get quoted by AI tend to attract links, attention, and traffic over time — often more reliably than pages that ranked through bulk.

A useful mindset

A page is the right length when every paragraph earns its place. Word counts measure effort, not value.

If you’ve written something tight, you’ve done well. If you’ve written something long that’s also tight, you’ve done well too. The thing to avoid is writing something long because it’s long — and that’s the thing most SEO advice still implicitly encourages.


Coming up in the next module: Structured data, made simple. We leave the writing patterns behind and move into schema markup — the structured language that tells AI systems exactly what each piece of content on your page actually is. Four lessons, focused entirely on what website owners need to understand without becoming developers.