Meta Tags Explained

Meta tags are still worth a few minutes of your time in 2026 — but only a handful of them. Roughly 90% of the tags that SEO sites were pushing in 2008 do nothing today, and search engines ignore most of the rest. This post keeps the original 2008 rundown below the fold and adds what actually matters now.

The short version for modern HTML

If you do nothing else, ship these five:

<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Page title</title>
<meta name="description" content="One compelling sentence. Google may use it as the snippet.">
<link rel="canonical" href="https://example.com/page">

• charset — technically an element attribute rather than a classic meta tag, but it must be within the first 1024 bytes of the document. Mismatched charsets are still a top cause of broken text.
• viewport — required for responsive layout; without it, mobile browsers render at ~980px and scale down.
• description — Google says it doesn’t affect ranking, but it often becomes the search snippet, which drives click-through. Google may rewrite it when it thinks it can do better.
• canonical — the modern replacement for half the “not recommended” tags below: it tells crawlers which URL is authoritative instead of relying on Meta Refresh hacks.
• Open Graph / Twitter cards (og:title, og:image, twitter:card) — the true successors of 2008-era metadata. Social platforms and chat apps ignore everything else when building link previews.

What happened to the rest

• keywords — Google hasn’t used it since 2009 (confirmed in the official Google Search Central blog, “Google does not use the keywords meta tag in web ranking”; Google’s current supported-tags documentation states it “has no effect on indexing and ranking at all”). Bing also ignores it. Filling it only hands your keyword research to competitors via View Source.
• robots — mostly redundant: pages are indexable by default. It still has one real use, noindex, for pages you can’t remove via robots.txt or HTTP headers. Since 2019 Google also supports max-snippet:, max-image-preview:, and unavailable_after: directives here.
• Meta Refresh (<meta http-equiv="refresh">) — W3C advises against it; it’s a hallmark of doorway pages and ad redirects. Use HTTP 301/302 or, in a pinch, JavaScript. Google treats short-delay refreshes as redirects, long ones as a crawl hint only.
• revisit-after — never honored by any major search engine. Dead since the 90s.
• Content-Script-Type / Content-Style-Type / Content-Language — the HTML spec dropped them; HTML5 removed http-equiv="Content-Language" entirely (use the lang attribute instead). Script and style defaults were never implemented by browsers.
• MS Smart Tags, MSN No-ODP, VW96.ObjectType, Rating, Generator — vendor-era relics. The only one with residual value is generator, which some site-builder frameworks emit for fingerprinting.

One modern addition worth knowing: in 2021 HTML gained <meta name="color-scheme" content="dark light">, which tells the browser which color schemes the page supports and prevents white flashes on dark-mode pages.

2008 original

The original rundown, preserved from October 2008 (opinions were those of the era’s SEO guides):

Recommended META-Tags: Meta Content Language (non-US English only), Meta Content Type, Meta Description. Optional: Meta Abstract, Meta Author, Meta Copyright, Meta Designer, Meta Google, Meta Keywords, Meta MSN (No ODP), Meta Title. Not recommended: Meta Content Script Type, Meta Content Style Type, Meta Distribution, Meta Expires, Meta Generator, Meta MS Smart Tags, Meta Pragma No-Cache, Meta Publisher, Meta Rating, Meta Refresh, Meta Reply-To, Meta Resource Type, Meta Revisit After, Meta Robots, Meta Set Cookie, Meta Subject, Meta VW96.ObjectType.

Meta Abstract — gives a short summary of the description, used primarily with academic papers; content usually 10 words or less. Example: <META NAME="Abstract" CONTENT="Short description of page">. Optional; it will not assist you with the major search engines. For highly specialized content it may allow search engines that are part of your field of expertise to index your website correctly; at the time it was not part of Google, Yahoo!, and MSN algorithms.

Meta Author — declares the author of the HTML or XML document. Example: <META NAME="Author" CONTENT="George Costanza, gcostanza@vandalayindustries.com">. If you use it, use the author’s first and last name. Including the email address is not recommended — site scrapers harvest addresses for spam; a contact form is safer. Google, Yahoo!, or MSN did not index the tag, so it wouldn’t help ranking, but it was recognized as part of the “Meta Tag Standard”. Place it in the HEAD section, following your Meta Description and Meta Keyword tags. A practical use beyond search: identifying which author wrote which pages, and proving who originally wrote a page when other sites rip off your content and neglect to strip the tag.

Meta Content Language — <META HTTP-EQUIV="Content-Language" CONTENT="en-GB">, for pages written in non-US English; robots may use it to categorize by language.

Meta Content Script Type — <META HTTP-EQUIV="Content-Script-Type" CONTENT="text/javascript">. Recommendation then, as now: do not use — browsers detect scripts on their own.

Meta Content Style Type — <META HTTP-EQUIV="Content-Style-Type" CONTENT="text/css">. Same verdict: do not use.

Meta Content Type — the one 2008 recommendation that still holds, in its modern form <meta charset="utf-8">. Declaring the charset has been mandatory in HTML5 since 2014; failure to declare it can cause display problems where, for instance, the document uses UTF-8 punctuation characters but is displayed in an ISO or ASCII charset.