Meta Tags for SEO: The Only Ones That Still Matter

Open the source of any web page and you will find a block of meta tags near the top — some essential, some harmless, and some pure cargo cult from 2004. Search engines ignore most of them now, but the handful that still matter punch far above their weight: they decide your page's title in search results, its preview on social media, and whether Google treats it as the canonical copy. Here is the short list worth your time.

The title tag: your single most important tag

The title element is not technically a meta tag, but it belongs at the top of this list because it is the strongest on-page signal you control. It becomes the clickable headline in search results, the label on the browser tab, and the default text when someone bookmarks or shares the page.

Keep titles around 50–60 characters so they do not get truncated in search results, put the most important words first, and make every page's title unique. A site where every page is titled "Home" or "Products" is telling search engines that all its pages are about the same thing. Front-load the topic, add the brand at the end if there is room, and write for a human scanning a results page — because that human is the one who decides whether to click.

The meta description: not a ranking factor, still essential

The meta description does not directly affect rankings — Google has said so for years — but it often becomes the snippet of text shown under your title in search results, which means it directly affects click-through rate. A good description is 150–160 characters, summarizes what the page delivers, and includes the search phrase naturally, since matching terms appear bolded in the snippet.

Two rules matter here. First, write a unique description for every page; a duplicated or missing description lets Google write its own snippet from page text, which is sometimes fine and often worse. Second, treat the description as ad copy, not keyword stuffing: its job is to convince a searcher that this result answers their question, not to repeat the keyword five times.

Open Graph and Twitter Cards: controlling social previews

When a link is shared on Facebook, LinkedIn, Slack, or X, the preview card comes from Open Graph tags, not from the page content. The four that matter are og:title, og:description, og:image, and og:url. Without them, the platform guesses — and it often guesses the wrong image or a navigation fragment as the description.

For the image, use 1200 x 630 pixels, keep text out of the edges where cropping happens, and make sure the image actually exists at the URL you declare — a broken og:image produces a preview with no image at all. Twitter Cards (twitter:card, twitter:title, twitter:description, twitter:image) used to be separate, but X now falls back to Open Graph tags when Twitter-specific ones are missing, so a complete Open Graph set covers both. Set twitter:card to summary_large_image if you want the big-image layout.

The tags that handle technical correctness

• Viewport — required for mobile: width=device-width, initial-scale=1. Without it, mobile browsers render the desktop layout shrunk to fit, and mobile usability suffers.
• Charset — declare UTF-8 first in the head so the browser interprets characters correctly. A missing or late charset is how accented characters turn into mojibake.
• Canonical — tells search engines which URL is the official copy when the same content is reachable at several addresses (with or without trailing slash, with tracking parameters, HTTP vs HTTPS). It prevents duplicate-content dilution without deleting the alternate URLs.
• Robots — usually unnecessary, since the default is index and follow. Use it only when you need to block a page (noindex) or stop link-following (nofollow), and remember it is a request, not enforcement.

These four are set-and-forget on most sites, but they are also the ones that cause the strangest bugs when missing — a page that looks fine on desktop but broken on mobile, or two URLs splitting one page's ranking between them.

The tags you can safely skip

The meta keywords tag has been ignored by Google since 2009, and Bing treats it mainly as a spam signal — leave it out entirely. The revisit-after and expires tags were never honored by major search engines. Geo tags (geo.region, geo.placename) were a brief local-SEO fad that no major engine uses. And author, copyright, and generator tags are harmless but invisible to search: include them only if you have an internal reason, never as an SEO tactic.

Worked example: a page before and after

Consider a recipe page with no meta work: its title is "Page 7," it has no description, and its social preview pulls a random photo from the sidebar. In search it looks generic, on social it looks broken, and it competes with its own print-friendly URL for ranking.

After thirty minutes of meta work, the same page has the title "Easy Weeknight Pasta With Lemon and Parmesan," a 155-character description promising a 20-minute dinner, Open Graph tags pointing at a proper 1200 x 630 food photo, and a canonical tag pointing at the main URL. Nothing about the recipe changed — but the search result is now clickable, the social card is now appetizing, and the ranking signal is consolidated on one URL. That is the entire return on investment of meta tags: small effort, applied to the handful that matter, compounding across every page.

Validating your work

Before publishing, check the rendered result, not just the source. Search engines and social platforms cache previews aggressively, so use their debuggers — Facebook's Sharing Debugger, LinkedIn's Post Inspector, and X's Card Validator — to force a fresh fetch and see exactly what users will see. For search, Google Search Console shows which pages have duplicate titles or missing descriptions across the whole site, which turns a one-page fix into a site-wide checklist.

Frequently asked questions