Findrix
Technical AEO & Site Infrastructure

Heading Hierarchy

Heading hierarchy is the ordered structure of H1 to H6 tags on a page. Learn the rules and why AI engines lift content by heading.

TL;DR

Heading hierarchy is the ordered structure formed by H1 through H6 tags on a page. The H1 states what the page is about, H2 tags divide it into major sections, and H3 tags subdivide those, producing an outline a machine can read without interpreting a single sentence. That outline is how most parsers decide which part of your page answers a given question.

Why heading hierarchy matters

Engines extracting an answer need to identify a passage boundary. Headings supply those boundaries explicitly, so content under a clear heading is easier to lift and attribute.

How to structure headings

Use one H1 per page

It should state the page topic and closely match the title tag without being identical to it.

Never skip levels

H1 to H2 to H3. Jumping from H1 to H3 breaks the outline for parsers and screen readers alike.

Write headings as questions where natural

A heading matching how someone asks the question makes the section beneath it an obvious answer.

Make sections self-contained

Content under a heading should make sense when extracted alone, without depending on the paragraph above it.

Choose level by meaning, not size

If a heading looks too large, fix it in CSS. Picking H4 for visual reasons corrupts the structure.

Keep them descriptive

A heading reading Overview tells a parser nothing. One naming the actual subject tells it everything.

Heading hierarchy vs. title tag

Heading hierarchy: On-page structure. It organises content for readers and marks extractable sections for parsers.

Title tag: Metadata in the head. It names the page in search results and browser tabs and never appears in the page body.

They should agree without being identical. The title tag can carry brand and keyword framing for the results page; the H1 can read more naturally for someone who has already arrived.

Clean structure only pays off if crawlers can fetch the page and read it in the raw HTML. Findrix runs a 31-point technical audit covering rendering, structure and crawler access, then tracks how often seven AI engines cite you against named competitors. Every gap comes with the fix already written: technical, content and off-site. The audit is free, takes about a minute, and requires no signup.

Common heading mistakes

Why headings matter more in AI answers

Traditional search returned a link, and the reader arrived and found the relevant part themselves. AI answers extract a passage and present it directly, which means the engine has to decide where the relevant passage begins and ends before it can quote anything.

Headings are the strongest available signal for that boundary. A question-shaped heading followed by a direct answer is close to the ideal unit for extraction, and it is why well-structured documentation gets cited far more often than equally accurate prose.

The takeaway

Audit your important pages for headings that describe rather than label, and make sure the paragraph beneath each one answers it without requiring anything above.

Frequently asked questions

Can a page have more than one H1?

HTML5 permits it, and engines handle it. One H1 is still the better practice, because it states unambiguously what the page is about and keeps the outline simple for both parsers and screen readers.

Does heading hierarchy affect rankings?

Indirectly. Headings help engines understand structure and identify relevant passages, which affects snippet and citation selection more than ranking position. The accessibility benefit is direct and independent of search.

Should headings contain keywords?

They should describe the section accurately, which usually means the relevant terms appear naturally. Forcing keywords into every heading produces awkward labels that serve neither readers nor parsers.

← Back to the glossary