Heading Tags for SEO: How H1, H2 and H3 Actually Work

An audit told you to fix your heading tags, handed you a checklist, and moved on to the next item, without telling you whether any of it would actually change your rankings. That's the honest gap in most heading tags SEO advice: plenty of checklists, very little about which parts genuinely move the needle and which are pure hygiene. Google's own documentation, and its own engineers, have said enough in public that you don't have to guess. Here's what heading tags actually do, what breaks when you get them wrong, and, just as importantly, what fixing them will not do for you.

Heading tags, H1 through H6, are the HTML markup that tells a browser, a screen reader, and a search engine which piece of text is a title and which pieces are subsections beneath it. An H1 marks the page's main heading. H2s mark the major sections under that H1. H3s mark subsections within an H2, and so on down to H6, which almost no small business page ever genuinely needs. Visually, you could make any text look like a heading with bold formatting and a bigger font. Structurally, only text wrapped in an actual heading tag does the job search engines and assistive technology rely on, which is why a page that looks correctly structured to the eye can still be a mess underneath it, invisible until someone actually checks the markup.

You're thinking a perfectly optimised heading hierarchy will move your rankings. Google's own John Mueller has been direct about this in public: headings are great for SEO and accessibility, in his words, but they're not going to make or break your site's rankings. He's also said a site will rank perfectly fine with no H1 tags or with five of them, heading tags are not, on their own, a ranking factor the way genuine topical relevance or a working technical foundation are. Keywords in heading tags stopped being a meaningful ranking signal on their own more than fifteen years ago, per Mueller, even though the habit of stuffing them in persists.

What headings do reliably do is help Google understand what a specific piece of text or an image on your page is actually about, and Google's own Search Central documentation names heading elements specifically as one of the sources it draws on when generating the title link shown in search results, alongside your page's title tag and other prominent text. That's a real, functional role. It's just not the same as a ranking lever, and treating it as one is how businesses end up reshuffling headings for months while the actual problem, thin content or a weak keyword match, goes untouched.

Three rules cover almost everything about heading structure that's actually worth fixing:

The accessibility case is the one most heading-tag advice skips entirely, and it's arguably the more important one. The Web Content Accessibility Guidelines don't technically fail a page under Success Criterion 2.4.6 purely for skipping a heading level, but skipping levels breaks the logical structure screen reader users rely on to navigate a page by jumping between headings, which is the practical function WCAG's Success Criterion 1.3.1 (Info and Relationships) is actually protecting. A sighted visitor skims your page with their eyes. A screen reader user skims it by heading level. Get the hierarchy wrong and the page becomes functionally harder to navigate for a real share of your visitors, independent of anything Google ever measures.

Two H1 tags on one page is the most common mistake, usually one from the page template and a second buried in a hero banner or CMS block nobody remembers is still there. Neither Google nor a visitor typically notices the duplication outright, but it usually signals a templating problem worth fixing anyway, since the moment the wrong text becomes the visible H1, so does whatever page title Google chooses to display. The second most common: heading text stuffed with the exact keyword phrase in every single H2 on the page, a leftover habit from an era when Google genuinely counted keyword occurrences in headings as a ranking signal. It doesn't work that way anymore, and to a human reader it reads exactly as clumsy as it sounds, which costs trust with the person you're actually trying to convert.

Here's the section most agency content skips, because it argues against selling you an hour of heading cleanup. If your rankings haven't moved and headings are the only thing that's changed, you're very likely optimising the wrong layer. A technically perfect heading hierarchy on a page that doesn't answer the query a real person is searching for will not outrank a messier page that actually does. Headings are structure, not substance, they help Google and your readers understand content that's already good, they don't create demand for content that isn't. If you've fixed your heading tags and traffic still hasn't moved after a couple of months, the real problem is almost certainly keyword targeting, content depth, or authority, not the H2 you just renamed.

The one genuine exception: a page with badly broken heading structure, no H1 at all, or heading tags entirely absent, is a real accessibility and clarity problem worth fixing regardless of whether it moves a single ranking position, because real people using screen readers are navigating your page by exactly this structure today, not at some future point once an algorithm rewards it.

AI answer engines, ChatGPT, Perplexity, Google's AI Overviews, work by pulling specific, extractable passages out of a page rather than reading it start to finish the way a person does. A clear heading followed immediately by a direct, self-contained answer is exactly the shape those systems tend to extract most reliably; a wall of unstructured prose with no heading markers is harder to parse into a discrete, quotable fact. This is a widely observed pattern in how well-cited content tends to be structured, not a formally published algorithm rule from any of these companies, so treat it as good practice rather than a guaranteed mechanism. Blackstone Media's own site takes the underlying access question seriously regardless of heading structure: robots.txt explicitly allows GPTBot, ClaudeBot, PerplexityBot, and OAI-SearchBot, and a llms.txt file sits at the site root, the access layer AI crawlers need before heading structure or anything else in the content can be read at all.

If a full technical read of your site's headings, crawlability, and structure sounds like more time than you have this week, that's exactly the kind of audit our SEO and GEO agency team runs as a matter of course, not an upsell, and we'll tell you plainly which fixes will move traffic and which are just hygiene. For the fuller crawlability and structure checklist beyond headings alone, see our guide to SEO friendliness. Or contact the Blackstone Media team directly.

Do heading tags (H1, H2, H3) directly affect Google rankings?

Not directly, according to Google's own John Mueller, who has said publicly that a site will rank fine with no H1 tags or with five of them. Headings help Google understand what a page and its images are about, and factor into how Google generates the title link shown in search results, a real, functional role, just not a ranking lever on their own.

How many H1 tags should a page have?

One. It should state the page's actual topic, not a repeated business name or a vague slogan. Having two H1s rarely causes a visible penalty on its own, but it's usually a sign of a templating issue worth fixing, and it can confuse which text Google treats as the page's main heading when generating a title link.

Does skipping a heading level, like going from H2 straight to H4, hurt my SEO?

It's not a documented Google ranking factor, but it does break the logical structure screen reader users rely on to navigate your page by heading, a genuine accessibility issue under WCAG's Info and Relationships criterion (1.3.1) independent of anything Google measures. Fix it for your actual visitors, not for an algorithm.

Should I put my target keyword in every heading on the page?

No. Keyword-stuffed headings were a weak, and now outdated, ranking tactic from over fifteen years ago, per Google's own John Mueller. Write headings that describe what the section actually covers; include your keyword naturally where it genuinely fits, not in every single heading on the page.

Do heading tags matter for AI search tools like ChatGPT or Google AI Overviews?

Clear, direct headings followed immediately by a self-contained answer are easier for AI systems to extract as a quotable passage than unstructured prose is, a widely observed pattern in how AI-cited content tends to be structured, not a formally published algorithm rule. It's good practice regardless of whether you're optimising for AI citation specifically.

How do I check my own site's heading structure?

Use your browser's Inspect or View Page Source tool and search for every <h1> and <h2> tag on the page, or run a free heading-structure or accessibility checker, which will flag skipped levels and missing or duplicate H1s directly rather than requiring you to read the raw HTML by eye.