What Is an H1 Tag in SEO?
Headings get treated as a search device. They are a document structure device that search engines happen to read. Framed that way the advice becomes obvious rather than arbitrary.
Headings Are Structure, Not Decoration
A heading describes the shape of a document: this is a section, this sits inside it, this begins something new. What size it appears at is a separate decision entirely. Confusing the two causes almost everything that goes wrong here.
What a heading declares. Position.
That a new part of the document starts here and where it sits relative to everything around it. That statement is made regardless of how the text is styled.
Why the confusion arises. Editors show size.
Most editing tools present heading levels as sizes, so choosing one looks like a formatting decision. It is a structural declaration that happens to have a default appearance attached.
What follows from the reframe. The advice writes itself.
Once headings are understood as structure, questions about skipping levels and using several of one kind answer themselves by asking what the document actually looks like.
Who reads the structure. More than search engines.
Browsers, assistive software, anything converting a page to another format and anybody scanning rather than reading. Search engines are one audience among several, which block eight develops.
What An H1 Is
The heading that names the whole page. Everything else on the page sits underneath it, which is the entire meaning of the level rather than a claim about importance.
What it should contain. The subject.
What this page is about, in the words a reader would use. It is the answer to somebody asking what they are looking at, which is a lower bar than it sounds and frequently missed.
Where it appears. On the page.
Visible to anybody reading, usually at the top. That distinguishes it from the title, which is written for results and rarely seen on the page itself.
What happens without one. Less than people fear.
A page with no H1 is not penalised. It is harder to navigate and its structure is less clear, which are the real costs rather than a search one.
What platforms usually do. Provide it automatically.
Most systems make the page title an H1 without being asked, which is the right behaviour. Problems arise where somebody overrides it for appearance.
Title Tag Against H1
One appears in search results and one appears on the page. They may reasonably differ. Treating a difference as an error is the commonest misunderstanding about both.
Why they serve different readers. Different moments.
The title is read by somebody deciding whether to visit, alongside competing results. The heading is read by somebody who has already arrived and wants confirmation they are in the right place.
What that permits. Different phrasing.
A title can be shorter and more direct because it competes for attention. A heading can be fuller because the reader is already committed.
A worked example from this site. Deliberate.
Our hosting guide carries a short statement of the subject as its title and a question as its heading. Nobody made a mistake: one is written for a results page and one for somebody who arrived.
The sensible relationship. Same subject, different shape.
They should be recognisably about the same thing. A title and heading describing different subjects means one of them is wrong, which the meta tags pillar covers as a rewrite trigger.
Only One H1
The older rule has relaxed. One remains the sensible default because a document has one subject. Using more is not a penalty, which Google has said directly.
What Google has stated. Repeatedly.
In a video published in 2019, Google's John Mueller said their systems have no problem with multiple H1 headings, that it is a common pattern and that headings are used to understand the context of parts of a page.
What he added later. No bound either way.
In a subsequent office hours session he stated there is no upper or lower limit on H1 elements, adding that a site will perform perfectly well with none or with several.
Why one is still the default. Documents have one subject.
If a page needs two top-level headings, it is usually two pages. That is a structural observation rather than a rule. It connects directly to one subject per page.
Why audit software disagrees. Older conventions.
Automated reports routinely flag multiple H1 elements as an error. That reflects a convention rather than a published position, which is worth knowing when a report arrives.
H2 And H3 Tags
Levels describe nesting rather than importance. An H3 is not a less significant point than an H2, it is a point that sits inside one, which is a different claim entirely.
What each level means. Containment.
A second-level heading opens a major section of the page. A third-level heading opens something within that section. The relationship is where each sits, not how much it matters.
Why importance is the wrong reading. It produces nonsense.
Treating levels as a ranking of significance leads people to promote a crucial point to a higher level regardless of where it belongs, which breaks the outline it was meant to clarify.
How deep to go. Rarely far.
Most business pages need two levels and occasionally three. A page reaching further usually has a scope problem rather than a structure problem.
The test for a new section. Would it survive alone.
A heading should open something with enough substance to be worth naming. A heading over two sentences is decoration wearing a structural label.
Structuring Headings
Somebody should be able to follow the argument of a page from its headings alone. A page failing that test has a structure problem rather than a heading problem, which is a more useful diagnosis.
How to apply it. Read only the headings.
Skip everything else and see whether the sequence makes sense. If it reads as a coherent outline, the page is structured. If it reads as disconnected labels, it is not.
What failure looks like. Headings that describe nothing.
Single words, vague nouns and phrases that could sit above any section. Those tell a scanning reader nothing and force them to read everything to find one thing.
Why it is a structure problem. The headings are a symptom.
Vague headings usually mean the page has no argument to describe. Rewriting them without fixing the underlying order produces better labels on the same confusion.
What good headings do. Carry meaning alone.
Each one states something rather than naming a topic. That is the difference between a label and a sentence somebody could act on.
Do Not Skip Levels For Appearance
Choosing a heading level because of how it looks breaks the structure. Styling should be handled separately. This is an accessibility failure as much as a search one.
How it happens. Reasonably.
Somebody wants a smaller heading, so they use a lower level. The text looks right and the document now claims a section sits inside something that does not exist.
What it produces. A broken outline.
Anything reading the structure sees gaps and misplacements. The page looks correct to somebody with the whole screen in front of them and is incoherent to anything reading it in sequence.
The correct fix. Change the styling.
Heading levels and their appearance are set separately. Wanting a smaller heading is a styling request. Doing it properly costs a developer minutes.
Why it survives. Nothing complains.
No error is produced, the page looks intentional and nobody sighted encounters a problem. It is found by looking deliberately or by somebody using assistive software.
Accessibility Is The Better Argument
Screen readers navigate by headings. A broken structure genuinely disadvantages people. That reason holds whatever any search engine does with headings.
How the navigation works. Jumping between them.
Assistive software can move through a page heading by heading, which is how somebody finds the part they want without hearing everything before it. The WebAIM screen reader survey finds this is how most such users navigate.
What a broken structure does to that. Removes the map.
Skipped levels and headings used for size produce an outline that misrepresents the page. Somebody navigating by it is being given wrong directions rather than none.
Why this argument is stronger. It does not depend on a claim.
Ranking effects of heading structure are modest and arguable. The effect on somebody using a screen reader is direct, immediate and not in dispute.
Where else this applies. Twice more here.
Alternative text on images and general page usability follow the same pattern. Where a search claim is not established, the accessibility reason is what makes the advice defensible.
Stuffing Headings Does Not Work
Repeating a phrase across every heading is visible, counterproductive and reads badly to anybody. It is also one of the documented reasons a title gets replaced in results.
What it looks like. The same words, repeatedly.
Every heading containing the service and the town, arranged slightly differently. Obvious to a reader within seconds and impossible to unsee once noticed.
Why it fails as a technique. Nothing rewards it.
Repetition in headings is not a signal anything acts on, so the practice was built on a model of how search worked rather than on how it works.
What it costs. Credibility.
A page reading as written for software rather than for a person undermines the trust the page exists to build. Our risky practices material covers the wider pattern.
What to do instead. Describe the section.
A heading stating what that part of the page covers will contain the relevant language naturally, because it is describing the thing.
Our Own Standard
Every block on our guides carries a short descriptive line above a Title Case heading. That is a deliberate arrangement. The reasoning is worth setting out, since you can see it on this page.
What the arrangement is. Two parts.
A small line describing what the block does, then a short heading naming the subject. The first is styling rather than structure, so it does not appear in the outline.
Why the headings are short. They are read alone.
Somebody scanning sees only the headings, so each has to work without its surroundings. Long headings read poorly in that sequence.
What the descriptive line adds. Context without length.
It carries the nuance a short heading cannot, for somebody reading properly, while keeping the outline clean for somebody scanning.
Why we are showing our own working. It is checkable.
You can read this page's headings in sequence and judge whether the structure holds. That is more useful than asserting a standard. Optimising existing content applies the same test.
Read only your
headings. Does it
still make sense?
If the sequence reads as a coherent outline, the page is structured. If it reads as disconnected labels, the headings are a symptom: the page has no argument to describe, so better labels on the same confusion will not fix it.
How we structure a page:
If an audit flagged your headings as errors, ask what published position it is applying.
Every guide.
One practice.
What a page says about itself, headings, keywords, search intent, images, page types, user behaviour, credibility and how to review an existing site.
