Meta Tags for SEO

TL;DR — Meta tags are bits of HTML in the head of your page that tell search engines and browsers things about the page. Most of them don’t matter for SEO. The three that do: the title (the clickable headline in results), the meta description (the blurb under it), and the robots tag (which can tell Google not to index a page). Meta keywords? Dead — Google ignores them completely.

Evidence for this claim The HTML meta element represents metadata that other head elements cannot express. Scope: HTML semantics, independent of whether a particular search engine uses a given name value. Confidence: high · Verified: 2026-07-14 MDN: meta element Evidence for this claim Google documents a defined set of meta tags it understands and ignores unsupported tags. Scope: Google Search's supported meta-tag behavior. Confidence: high · Verified: 2026-07-14 Google Search Central: Meta tags and attributes

What meta tags are

When you look at a web page, you see the body — headings, paragraphs, images. Above all of that, in a section called the head, sits metadata: information about the page that visitors don’t see but search engines and browsers read. Meta tags live there.

The internet is full of “SEO meta tags” checklists with a dozen entries. Ignore most of them. In 2026, only a few head elements actually do anything for search:

• The title — the headline that becomes your clickable link in search results. (Technically it’s the title element, not a “meta” tag, but it belongs here and it’s the most important one.)
• The meta description — the short blurb shown under the title. It doesn’t help you rank, but a good one gets more people to click.
• The meta robots tag — how you tell search engines not to index a page, or not to show a snippet for it.

The tags that do nothing

• Meta keywords — you list keywords, Google ignores them entirely. Filling this in is pure wasted effort, and it’s been that way since 2009.
• Author, revisit-after, and similar — never used by Google.

If a “must-have meta tags” article tells you to add these, close the tab.

The one mistake to avoid

People try to hide a page from Google two ways at once: they block it in robots.txt and add a noindex tag. That backfires. If you block the page, Google never reads it — so it never sees the noindex, and the page can linger in search. If you want a page out of Google, let it be crawled and add noindex. Don’t block it.

Want the full map — every tag, where it lives, and which ones control indexing vs. snippets? Switch to the Advanced tab.

TL;DR — Group head elements by job, not by name. Appearance: the title element (a primary — not only — source of the SERP title link) and the meta description (CTR copy, not a ranking factor). Indexing & serving: the robots meta tag (noindex, nofollow, nosnippet, max-snippet…), mirrored by the X-Robots-Tag HTTP header for non-HTML files and data-nosnippet at element level. Link-level: rel="nofollow", sponsored, ugc on <a> tags — hints, not directives. Browser-only: charset, viewport, refresh. Ignored: keywords, author, revisit-after. The load-bearing rule: robots.txt is crawl control, the robots meta tag is index control — block a page and Google never sees your noindex.

Evidence for this claim The HTML meta element represents metadata that other head elements cannot express. Scope: HTML semantics, independent of whether a particular search engine uses a given name value. Confidence: high · Verified: 2026-07-14 MDN: meta element Evidence for this claim Google documents a defined set of meta tags it understands and ignores unsupported tags. Scope: Google Search's supported meta-tag behavior. Confidence: high · Verified: 2026-07-14 Google Search Central: Meta tags and attributes

What counts as a meta tag — and the title-tag caveat

Meta tags are elements in the head that pass metadata about the page to search engines and browsers. The classic ones use the meta element. But the single most important head element for SEO — the title element — is technically not a meta tag. I’m covering it here anyway, because every “meta tags for SEO” search wants it, and because grouping head elements by the job they do is far more useful than splitting hairs over which ones are literally <meta>.

So here’s the map I want you to keep in your head.

The map — every head signal by its job

1. Appearance (matters for SEO).

• The title element — becomes your title link in results. It is a light ranking factor, and it’s the primary, but not the only, source of the SERP title. Google: it “uses a number of different sources to automatically determine the title link,” including your title, the visible H1, prominent text, og:title, and anchor text. So it can rewrite yours. Ahrefs’ large-scale research found Google rewrites titles roughly a third of the time, often when they’re too long or boilerplate. Aim for ~60 characters / under ~600px.
• The meta description — the snippet copy. It is not a ranking factor (more on that below), but it’s the pitch that earns the click — write it as conversion copy, not keyword copy. Google rewrites or ignores it the majority of the time (Ahrefs’ data puts the rewrite rate around 62.78%), so spend the effort on high-value, high-traffic pages and don’t sweat it on the long tail.

2. Indexing & serving control (the robots family). This is one set of directives delivered in up to three places:

• The robots meta tag in the head — noindex, nofollow, none (= noindex, nofollow), plus snippet controls nosnippet, max-snippet, max-image-preview, max-video-preview, and noimageindex, notranslate, unavailable_after, indexifembedded. Target a specific bot with name="googlebot" instead of name="robots".
• The X-Robots-Tag HTTP header — the same directives, set server-side in the response header. This is how you noindex a PDF, an image, or any non-HTML file that has no head to put a meta tag in. Underused, and a real differentiator.
• The data-nosnippet attribute — element-level snippet control. Wrap a span, div, or section and that text is excluded from the snippet without hiding the whole thing. It’s the only sub-page snippet control.

A 2026-current detail worth flagging: max-snippet now also caps how much of your content can feed AI Overviews and AI Mode — Google’s docs say so explicitly. And when directives conflict, the more restrictive rule wins (nosnippet beats max-snippet:50).

3. Link-level signals (on <a> tags, not in the head).

• rel="sponsored" — paid links and ads.
• rel="ugc" — user-generated content (comments, forum posts).
• rel="nofollow" — everything else you don’t want to vouch for. Since 2019, Google treats these three as hints, not strict directives, for ranking — and flagged links can still be discovered and crawled through other paths. Don’t confuse the link-level rel="nofollow" (one link) with the page-level robots meta nofollow (every link on the page) — different scope, different mechanism.

4. Browser-only (not SEO directives). charset/content-type and viewport are for rendering, not ranking. The meta-refresh tag — Google calls it a tag that “sends the user to a new URL after a certain amount of time” — is a browser convenience; for an SEO redirect use a real server-side 301, not a refresh.

5. Ignored / dead. keywords (“no effect on indexing and ranking at all”), nositelinkssearchbox (now a no-op), and author/revisit-after (never used by Google). Stop cargo-culting these.

Where each tag lives — three locations

This is the other axis that trips developers up:

• HTML head — the title element, meta description, robots meta tag, and the browser tags.
• HTTP response header — X-Robots-Tag, carrying the same robots directives, set server-side, working for any file type.
• On the element / link — data-nosnippet on a span/div/section, and the rel values on <a> tags.

Crawling vs. indexing vs. serving — why robots.txt can silence your meta tags

This is the most important concept on the page, so I’ll be blunt about it. These are different layers:

• robots.txt = crawl control. It decides whether Google fetches the URL at all.
• robots meta tag / X-Robots-Tag = index + serve control. They decide whether a crawled page gets indexed and how it’s shown.
• data-nosnippet = serve control only, at element granularity.

Here’s the gotcha, straight from Google’s docs: if a page is “disallowed from crawling through the robots.txt file, then any information about indexing or serving rules will not be found and will therefore be ignored.” Block the page and your noindex is never seen. The page can sit in the index with no snippet, indexed purely off links pointing to it.

I’ve tested the crawl side of this directly. In The Story of Blocking 2 High-Ranking Pages With Robots.txt I blocked two of our own ranking pages and watched what happened. My takeaway holds up: “Don’t block pages you want indexed. It hurts. Not as bad as you might think it does—but it still hurts.” The pages didn’t vanish — they kept ranking, just without their snippets and featured-snippet slots. Which is exactly the point: blocking is a crawl lever, not a deindex lever. To remove a page, allow crawling and add noindex.

The meta description is not a ranking factor

It’s worth stating flatly because the myth is so durable. John Mueller put it plainly: the meta description “is primarily used as a snippet in the search results page. And that’s not something that we would use for ranking.” It affects CTR, not position. The modern wrinkle: AI and LLM surfaces increasingly lift it as a ready-made summary, so a tight description earns you good representation in more than just the blue links now.

Two implementation gotchas that quietly break metadata

Invalid HTML in the head truncates everything after it. The <head> element may only contain title, meta, link, script, style, base, noscript, and template. Google’s docs are blunt about what happens if you break that: “If you use an invalid element in the head element, Google ignores any elements that appear after the invalid element.” The two invalid elements Google names explicitly are iframe and img. If a template accidentally drops one of those into the head — ahead of your meta description or robots tag — everything after it goes dark, even though the markup “looks” fine in the source. Audit tip: check your rendered head for a stray iframe/img before you go debug why a meta tag “isn’t working.”

Don’t let JavaScript inject or mutate meta tags if you can avoid it. Google’s own guidance: “We recommend that you avoid using JavaScript to inject or change meta tags whenever possible, and if you must use it, test your implementations thoroughly.” Server-rendered head tags are the stable default. But “inspect the rendered HTML” does not mean “the rendered value always wins.” Record each field by stage: title and description remain inputs to Google’s title-link/snippet selection; a raw noindex may stop rendering before JavaScript removes it; and Google says not to emit one canonical in source and change it later. The current JavaScript SEO guidance documents those field-specific differences.

Duplicate metadata is evidence, not a precedence rule. The HTML standard permits one title and one meta description per document. If a template emits several, show every raw and rendered value and flag the duplicate; do not claim that the first or last is Google’s guaranteed choice.

Robots meta in the body is a Google-specific exception. Valid, portable markup still puts the tag in <head>, but Google’s current robots documentation says Google Search reads and respects robots meta tags in the body. Treat body placement as a conformance/portability notice for Google, not proof the rule failed. Canonicals do not share this exception.

Where to go next

This hub is the map; each topic below is its own deep dive, and they all live in the on-page cluster sidebar. (For the full set, see the on-page cluster index.)

Appearance

• The title tag — your most important head element: how the title link is chosen, why Google rewrites it, and how long it should be.
• The meta description — writing snippet copy that earns the click, why it’s not a ranking factor, and how often Google rewrites it.

Indexing & serving control

• The robots meta tag (meta robots) — noindex, nofollow, none, and how to target specific bots.
• The X-Robots-Tag header — the same directives for PDFs, images, and other non-HTML files.
• noindex — how to actually remove a page from the index (and why it must stay crawlable to work).
• nofollow — the page-level directive vs. the link-level attribute, and what each one really does.
• rel=sponsored and rel=ugc — qualifying paid and user-generated links the way Google expects.
• max-snippet / max-image-preview — tuning how much text and how large an image Google can show, including the reach into AI Overviews.
• nosnippet / data-nosnippet — suppressing the snippet for a whole page or a single element.

Social sharing & platform tags

• Open Graph tags — og:title, og:image, and the rest of the protocol that controls how a link previews on Facebook, LinkedIn, Slack, and Discord.
• Twitter/X Card tags — the twitter:card family, and how X’s own volatile link-preview behavior has changed independently of your markup.
• Social sharing images — the og:image/twitter:image dimension, file-size, and absolute-URL rules that keep a shared link from showing a broken or stale preview.
• Favicon SEO — the <link rel="icon"> requirements for a favicon to actually show up next to your result in Google.

Legacy & rarely-useful tags

• The meta charset tag — why <meta charset="utf-8"> has to sit in the first 1024 bytes, and what happens when it doesn’t.
• The meta keywords tag — the tag Google has ignored since 2009, and why it’s still worth removing.
• rel=“alternate” — the general link-relation mechanism behind hreflang, feed autodiscovery, and the now-retired AMP annotation.

The layer that sits above all of these is crawling: a page has to be crawlable for any head directive to be seen, which is why robots.txt is its own topic and the prerequisite for everything here.

AI summary

A condensed take on the Advanced version:

• Meta tags = head metadata. Most “SEO meta tags” lists are padded; only a few do anything in 2026.
• Group by job, not name:
  • Appearance: the title element (a light ranking factor; the primary—not only—source of the SERP title link, which Google can rewrite ~a third of the time) and the meta description (CTR copy, not a ranking factor; rewritten ~62.78% of the time).
  • Indexing & serving: the robots meta tag (noindex, nofollow, none, nosnippet, max-snippet, max-image-preview…), mirrored by the X-Robots-Tag HTTP header for non-HTML files, and data-nosnippet at element level. max-snippet also caps input to AI Overviews / AI Mode; on conflict the more restrictive rule wins.
  • Link-level: rel="nofollow", sponsored, ugc on <a> tags — hints, not directives, since 2019.
  • Browser-only: charset, viewport, refresh (use a 301, not meta-refresh).
  • Ignored: keywords (“no effect on indexing and ranking at all”), author, revisit-after, nositelinkssearchbox.
• Three locations: HTML head, HTTP header (X-Robots-Tag), element/link (data-nosnippet, rel).
• The load-bearing rule: robots.txt (crawl) sits above the robots meta tag (index/serve). Block a page and Google never sees your noindex — to remove a page, allow crawling and add noindex. Patrick’s blocked-pages experiment: blocked pages mostly kept ranking, “but it still hurts.”
• Meta description ≠ ranking (Mueller: “not something that we would use for ranking”) — it drives CTR, and increasingly feeds AI summaries.
• Two implementation gotchas: an invalid head element (iframe, img) makes Google stop reading everything after it — audit the rendered head, not just the source. And Google says to avoid JavaScript-injected/mutated meta tags where possible; test the rendered output thoroughly if you can’t.

Official documentation

Primary-source documentation from the search engines.

Google

• Meta tags and HTML attributes that Google supports — the canonical list: description, robots, keywords (ignored), notranslate, refresh, verification, and the browser tags.
• Robots meta tag, data-nosnippet, and X-Robots-Tag specifications — every robots directive, where it lives, combining rules, and the crawl-vs-index dependency.
• Control your title links in search results — how the title link is chosen and why Google may rewrite it.
• Use valid HTML to specify page metadata — the valid <head> element list, and why an invalid element like iframe/img makes Google stop reading everything after it.
• Qualify your outbound links to Google — rel="sponsored", ugc, and nofollow.
• Evolving “nofollow” — new ways to identify the nature of links — the 2019 change that made the link attributes hints.

Bing / Microsoft

• Robots meta tags and attributes that Bing supports — Bing’s robots directives in both the meta tag and the X-Robots-Tag header, plus bingbot targeting. (Renders client-side; confirm specifics in a live browser.)

Quotes from the source

On-the-record statements from Google. Each link is a deep link that jumps to the quoted passage on the source page.

Google — what the tags do

• “Use this tag to provide a short description of the page. In some situations, this description is used in the snippet shown in search results.” — Google Search Central docs, on the meta description. Jump to quote
• “The meta-keyword tag is not used by Google Search, and it has no effect on indexing and ranking at all.” — Google Search Central docs. Jump to quote
• “A title link is the title of a search result on Google Search and other properties (for example, Google News) that links to the web page.” — Google Search Central docs. Jump to quote
• “Make sure every page on your site has a title specified in the title element.” (the docs reference the title element directly.) — Google Search Central docs. Jump to quote

Google — robots directives

• “In the case of conflicting robots rules, the more restrictive rule applies. For example, if a page has both max-snippet:50 and nosnippet rules…” — Google Search Central docs. Jump to quote
• “…disallowed from crawling through the robots.txt file, then any information about indexing or serving rules will not be found and will therefore be ignored.” — Google Search Central docs. The single most important rule for this hub. Jump to quote
• “If you use an invalid element in the head element, Google ignores any elements that appear after the invalid element.” — Google Search Central docs, on invalid <head> elements like iframe and img. Jump to quote
• “We recommend that you avoid using JavaScript to inject or change meta tags whenever possible, and if you must use it, test your implementations thoroughly.” — Google Search Central docs. Jump to quote

Google — rel attributes are hints

• “…are treated as hints about which links to consider or exclude…” — Danny Sullivan & Gary Illyes, Google Search Central blog (2019), on sponsored, ugc, and nofollow. Jump to quote
• “…the linked pages may be found through other means, such as sitemaps or links from other sites, and thus they may still be crawled.” — Google Search Central docs. Jump to quote

John Mueller, Google — meta description is not a ranking factor

• “So the meta description is primarily used as a snippet in the search results page. And that’s not something that we would use for ranking.” — John Mueller, SEO Office Hours (May 2022), via Search Engine Journal. Read the coverage

Patrick Stox — the crawl-vs-index proof

• “Don’t block pages you want indexed. It hurts. Not as bad as you might think it does—but it still hurts.” — from my experiment, The Story of Blocking 2 High-Ranking Pages With Robots.txt. Jump to quote

Meta tags cheat sheet

Every signal — where it lives and what it controls

Fast facts

• To remove a page from search: noindex (meta or X-Robots-Tag) on a crawlable page. Never noindex + a robots.txt block on the same URL.
• noindex PDFs and images with the X-Robots-Tag header — they have no head.
• On conflicting robots rules, the more restrictive one wins.
• Title link: primary source is the title element, but Google pulls from H1, og:title, anchor text, and visible text too — and rewrites ~33.4% of the time.
• Meta description: rewritten/ignored ~62.78% of the time — write it for high-value pages, skip the long tail.
• A stray iframe/img in the <head> makes Google stop reading everything after it — audit the rendered head, not just the source, if a tag “isn’t working.”
• Avoid JavaScript-injected/mutated meta tags where you can; if you can’t, test the rendered HTML, not just the component source.

The mental models

1. Group head elements by job, not by name. Appearance (title + description) · Indexing & serving (robots meta / X-Robots-Tag / data-nosnippet) · Link-level (rel on <a>) · Browser-only (charset, viewport, refresh) · Ignored (keywords, author). Once a tag is in the right bucket, what it does is obvious.

2. The three layers — crawl, index, serve.

• robots.txt = crawl (does Google fetch it?)
• robots meta / X-Robots-Tag = index + serve (does it get indexed, how is it shown?)
• data-nosnippet = serve only, at element level. Keep them separate and most “why is my page still in Google” confusion disappears.

3. The decision rule for removing a page. Want it gone from search? Allow crawling + noindex. Want bots to skip a URL space entirely (and don’t care about indexing)? robots.txt disallow. Never use disallow to deindex — Google never crawls the page to see your noindex.

4. Three locations, same directives. The robots directives aren’t tied to the head. The head meta tag, the X-Robots-Tag header, and (for snippets) data-nosnippet are three delivery mechanisms — pick by the file type and granularity you need, not by habit.

5. Directives vs. hints. The robots meta tag and X-Robots-Tag are directives — Google obeys them. The link-level rel values (nofollow, sponsored, ugc) are hints since 2019 — Google considers them, but flagged links may still be crawled. Don’t expect link-level rel to behave like a page-level noindex.

Test yourself: Meta Tags for SEO

Five quick questions on the on-page meta tags that matter. Pick an answer for each, then check.

Meta Tags

Meta tags are HTML elements in a page's head that pass metadata about the page to search engines and browsers. For SEO only a few matter — the title element, the meta description, and the robots meta tag — while meta keywords and most others are ignored.

Related: Title Tag, Meta Description, Robots Meta Tag, Noindex

Meta Tags

Meta tags are HTML elements that sit in the head of a page and carry metadata — information about the page that isn’t rendered in the body — for search engines and browsers. Most “meta tags for SEO” lists are bloated with tags that do nothing; in 2026 only a handful actually matter.

The mental model worth installing: head elements split by job. Appearance is the title element (technically not a meta tag, but the single most important head element — it becomes the clickable title link) and the meta description (the snippet copy that drives clicks, but not a ranking factor). Indexing and serving control lives in the robots meta tag (noindex, nofollow, snippet limits) — or the equivalent X-Robots-Tag HTTP header for non-HTML files like PDFs and images, or the element-level data-nosnippet attribute. Link-level signals (rel="nofollow", rel="sponsored", rel="ugc") live on <a> tags, not in the head.

The relationship to crawling is the part people get wrong: a page blocked in robots.txt is never crawled, so Google never sees a noindex or any other head directive. Crawl control (robots.txt) and index control (the robots meta tag) are different layers. And meta keywords are dead — Google ignores them entirely.

Related: Title Tag, Meta Description, Robots Meta Tag, Noindex

Changelog