Heading tags are the H1 to H6 HTML elements that give a page its structure. Use one H1 that states what the page is about. Use H2s for your main sections and H3s for the points inside them. Keep the levels in order, write every heading so it makes sense on its own, and work keywords in only where they fit naturally.
Key Takeaways
- One H1 per page, matching the page’s main topic
- H2s mark sections; H3s break those sections into parts. Never skip a level
- Headings are the cut lines AI systems use to chunk and cite your content
- Vague headings like “Overview” or “Tips” get skipped by readers and models alike
- More headings is not automatically better. Structure should follow the content, not the other way around
What Are Heading Tags in SEO?
Heading tags are HTML elements that label the sections of a page. They run from <h1> down to <h6>, and the number signals importance. H1 is the page title. H6 is the deepest sub-point most sites never need.
They also go by header tags, H tags, or HTML headings. Same thing.
Here’s what each level is for:
| Tag | Role | How often |
|---|---|---|
| H1 | The page’s main topic. The “cover title” | Once per page |
| H2 | Main sections of the page | 4 to 8 on a typical article |
| H3 | Sub-points inside an H2 section | As needed |
| H4-H6 | Deeper nesting for complex or technical docs | Rarely |
One thing people mix up constantly: a heading tag is not a title tag. Your title tag lives in the <head> and shows up in Googconstantly mix up is thatidentical word for wordle’s results. Your H1 lives on the page and is what a visitor reads first. They should cover the same topic without being word-for-word identical. If you want the full breakdown on the other half of that pair, our guide to writing better page titles covers it: title tag best practices.
Why Heading Tags Matter (and What the Data Actually Shows)
Google’s John Mueller has described heading elements as a “really strong signal” for understanding what a page covers. That’s about as direct as Google gets on any on-page element.
But here’s the part most guides leave out. SearchPilot ran a 50/50 split test on the Moz blog, whose post headlines were wrapped in H2 tags instead of H1s. Half the headlines were switched to H1s. After eight weeks, the change produced no statistically significant difference in organic traffic. Google worked out the topic either way.
So headings are not a ranking lever you pull. That’s the wrong mental model.
Headings don’t move rankings directly. They decide whether your content is findable, scannable, and quotable, which is what moves rankings.
Three things actually depend on them:
Scanning. Readers don’t read. They skim headings, decide whether the page has their answer, and bail if it doesn’t. A page with clear H2s holds people. A wall of text loses them.
Accessibility. Screen readers announce each heading and its level, and let people jump between sections with a keystroke. Broken hierarchy makes a page genuinely hard to navigate.
AI retrieval. This is the one that changed. LLMs don’t read your page top to bottom. They break it into chunks and score each one, and H2/H3 boundaries are the natural cut lines they use. A tight heading followed by a self-contained answer is an extractable chunk. A heading followed by three tangled ideas is not. If ChatGPT, Perplexity, or Google’s AI Overviews are going to cite you, they’re citing a section, not a page.
How to Optimize Heading Tags: 7 Best Practices
1. Use exactly one H1. Google has confirmed its systems handle multiple H1s fine, but that’s a reason not to panic, not a reason to add more. Two H1s tell a reader the page has two main topics. Keep one.
2. Put your primary keyword in the H1, near the front. Naturally. “Heading Tags in SEO: Best Practices” works. “Heading Tags SEO Header Tags H1 Tag Guide” does not, and Google reads that as stuffing.
3. Never skip a level. H1 to H2 to H3, in order. Don’t jump from an H2 to an H4 because the font size looked right. If you need smaller text, use CSS. Heading tags are structure, not styling.
4. Make every heading stand on its own. Pull any heading out of the article. Does it still tell you what that section contains? “Best Practices” fails the test. “Best Practices for API Rate Limiting” passes. Vague labels like “Overview,” “Key Points,” or “Things to Consider” give a model nothing to match a query against, so it skips the section.
5. One idea per heading. A heading like “Installation, Configuration, and Troubleshooting” forces both readers and retrieval systems to guess which topic the section is really about. Split it into three.
6. Front-load the answer under the heading. The first sentence below an H2 should answer the question the H2 raises. That’s the inverted pyramid, and it’s the single formatting habit that makes a section quotable.
7. Keep them short and consistent. Aim for under 60 characters. Pick sentence case or title case and use it site-wide.
Secondary keywords belong in H2s and H3s, but only where they describe the section honestly. A heading built to hold a keyword rather than describe content is obvious to a reader within about two seconds.
Heading Tag Mistakes That Kill Scannability
Running a directory means looking at a lot of other people’s pages. The submissions that get rejected for thin content share a pattern: every heading on the page is a keyword string rather than a description of what’s below it. The reader can’t tell the sections apart, so the page reads as one long block no matter how many H2s are technically in the HTML.
The most common problems, in rough order of how often they show up:
- Headings used for styling. An H3 chosen because the font size looked good, not because the content sits a level below an H2
- Theme-generated mess. Sidebars, widgets, and footers wrapped in H2s or H4s, wrecking the hierarchy outside your content area
- Duplicate H1s across pages. Every page titled with the company name tells Google nothing
- Stuffed headings. Keywords crammed in where they don’t belong, which Google recognizes and discounts
- Headings that don’t match the content underneath. A promise the section doesn’t keep, which sends readers straight back to the results page
Watch the theme problem in particular. Search engines and screen readers parse the whole page, not just your article body, so a template that wraps a sidebar widget in an H2 is quietly breaking your structure on every URL you publish.
When Not to Break Content Into More Headings
Most heading guides stop at “add more subheadings.” That advice has a limit, and in early 2026 Google’s Danny Sullivan said so out loud, warning publishers against slicing content into bite-sized chunks purely to please LLMs.
He has a point. The signal you want is structure that follows real content. Chopping a paragraph into four H3-labeled fragments so an AI can grab one doesn’t create clarity, it fakes it. And a page of thin fragments looks exactly like what it is.
Skip the extra headings when:
- The section is one idea. If a 90-word paragraph fully answers the question, an H3 splitting it in half serves nobody
- The piece depends on narrative flow. Opinion writing, case studies, and anything that builds to a conclusion get worse when you let people jump straight to the payoff
- You’re adding headings to hit a number. There is no target count. Some good articles have four H2s. Some have nine
The honest test: would a smart reader with no SEO knowledge find this heading useful? If yes, keep it. If it’s there for a crawler, delete it. Chunking works because it matches how people actually read online, not because it games retrieval.
How to Audit Your Site’s Headings in Under an Hour
Start with a crawl. Screaming Frog is free up to 500 URLs and gives you an H1 tab that flags every page with a missing, duplicate, or multiple H1. Sort by the issue column and you’ll have your fix list in about ten minutes. Sitebulb does the same with priority scoring layered on top. Either works, and both show up in most roundups of the SEO tools worth paying for.
Then spot-check by hand. Open your five highest-traffic pages, right-click a heading, and choose Inspect. The source panel jumps straight to that element so you can read the actual tag number instead of guessing from the font size.
Finally, run the outline test. Copy just the headings out of a page, in order, and read them as a list. Do they tell the article’s story without the body text? If not, that’s a rewrite, not a tagging problem. This is also the fastest way to spot the structural gaps that hurt navigation elsewhere on the site, which is where things like breadcrumb navigation start earning their keep.
[Image: Screaming Frog H1 tab showing missing and duplicate H1 warnings]
FAQ
Can a page have more than one H1?
Yes, and Google has said its systems handle it without trouble. But it confuses readers about what the page is actually about, and it’s still treated as poor practice by most auditing tools. Use one.
Do heading tags improve rankings?
Not directly, based on the split-test evidence. They improve rankings indirectly by making content easier to scan, easier to navigate with assistive tech, and easier for search and AI systems to extract answers from.
Should my H1 match my title tag?
They should cover the same topic but don’t need to be identical. The title tag is written to earn the click in search results. The H1 is written for someone who has already landed on the page.
Conclusion
Stop treating heading tags as a ranking checkbox. Write one clear H1, keep the levels in order, and make each heading specific enough to survive being read on its own. Then answer the question right underneath it. Run a crawl this week, pull the heading list off your top ten pages, and fix the ones that could belong to any article on the internet.
Sources referenced: SEO Sherpa’s analysis of the SearchPilot heading test and Search Engine Land’s reporting on content chunking.