Meta Tags for SEO: The 9 Head Tags Every Page Actually Needs

Why the head still matters

Everything in <head> is invisible to your visitors and decisive for everyone else — the search result, the LinkedIn card, the WhatsApp preview, the crawler deciding whether this page is a duplicate of another one. It is the only part of your page that gets read by machines before a human sees anything at all.

It is also where the most folklore has accumulated. SEO advice from fifteen years ago still circulates as fact, plugins add tags nobody asked for, and the result is a head section with thirty lines in it where nine would do the same job more reliably.

The useful mental model: three tags talk to search engines, four talk to social platforms, and two are housekeeping. Everything else is optional at best.

The nine tags that do the work

1. Title — the one that ranks

<title> is the single largest on-page ranking factor you fully control, and it is also the blue line somebody reads before deciding whether to click. Aim for 50–60 characters. Google truncates on pixel width rather than character count, so a title in capitals or full of wide letters gets cut earlier than the count suggests — put the words that matter first and you are safe either way.

Google also rewrites titles it judges unhelpful, which means a padded or vague one costs you twice: it fails to rank, and then it gets replaced with whatever Google scraped from your H1 anyway.

2. Meta description — the one that gets the click

Not a ranking factor. Google has said so repeatedly and it remains true. It is still one of the highest-leverage lines on the page, because it is your advertisement in the search results. 120–158 characters is the practical range: below 120 you are wasting space, above about 158 the tail becomes an ellipsis, and mobile cuts shorter still.

Leave it out and Google scrapes a fragment of body text instead. That fragment is rarely the sentence you would have chosen.

3. Canonical — the one that stops you competing with yourself

<link rel="canonical"> tells search engines which URL is the real one when the same content is reachable at several addresses. Every page should carry one pointing at itself, which costs one line and protects you the moment somebody shares your page with a tracking parameter attached. There is a full guide to canonicals here.

4. Robots — the one that decides if you exist

index, follow for anything you want in search results. Use noindex, follow for thank-you pages, internal search results and thin filter pages: it keeps them out of the index while still letting link equity flow through them. This is the tag most likely to be catastrophically wrong on a newly launched site, because staging environments ship with noindex and somebody forgets to remove it.

5. Viewport — the one that decides if it is usable

<meta name="viewport" content="width=device-width, initial-scale=1">. Without it, mobile browsers render your page at desktop width and shrink it, which is both unusable and a mobile-friendliness failure. It is one line and it has been standard for over a decade, and I still find pages without it.

6–8. Open Graph: title, description, image

og:title, og:description and og:image decide what your link looks like when it is pasted into Facebook, LinkedIn, WhatsApp, Slack or anywhere else. LinkedIn has no proprietary format of its own — it reads standard Open Graph tags, which is why getting these right covers most of the internet at once.

The image should be 1200 × 630 pixels and referenced with an absolute URL. A relative path is the most common reason a preview shows no image at all: the crawler resolves it against its own domain, not yours, and silently 404s.

9. twitter:card — the upgrade

X falls back to Open Graph when Twitter-specific tags are missing, so strictly you could skip this. It is worth three lines anyway, because twitter:card set to summary_large_image is what turns a small thumbnail into a full-width image card, and that difference is visible in click-through.

Four tags you can safely delete

Deleting these does not improve rankings. It makes the head readable, which means the next person to look at it can tell what is deliberate and what is inherited — and that does eventually improve rankings, because mistakes stop hiding in the noise.

What this looks like on real pages

A service page you want to rank

Title leads with the service and the place: 55 characters, the important words first. Description states the offer and who it is for in a sentence a stranger understands. Self-referencing canonical. index, follow. Open Graph pointing at an image that shows the actual work rather than a stock photograph of a laptop.

A blog post you want shared

Same three SEO tags, but the Open Graph title can differ from the search title — and usually should. What earns a click in a search result is specific and slightly dry; what earns a click in a feed is shorter and has a hook. There is no rule that says both have to be the same sentence, and treating them as separate jobs is one of the cheapest wins available.

A thank-you or confirmation page

noindex, follow, no Open Graph investment, and a canonical pointing at itself. These pages exist to be reached, not found. Leaving them indexable is how a “thanks, we will be in touch” page ends up as your top-ranking result for your own brand name.

A campaign landing page with tracking parameters

Self-referencing canonical to the clean URL, so the twelve UTM variants your campaign produces all consolidate onto one page. Pair the UTM builder with the canonical generator and this stops being something you have to remember.

Why use this generator

The counters run while you type. Most generators tell you the length after you have written the thing. Seeing the counter go amber at 61 characters changes the sentence you write, which is the entire point — length is a constraint on drafting, not a report card afterwards.

It produces the complete block, not just two lines. Title, description, canonical, robots, Open Graph and Twitter, formatted and ready to paste into <head>. Half the value of a generator is that it stops you forgetting the four tags you always forget.

Nothing is uploaded. It runs entirely in your browser — no account, no email gate, no logging. Unpublished page titles routinely give away a launch, and there is no reason for a third party to hold them.

It is honest about what it cannot do. No score, no traffic-light grade, no promise that a well-formed title will rank. Tags make a page technically capable of ranking. Content, intent and links decide whether it does. Anything claiming more than that is selling something.

A pre-publish routine

1. Write the title first, before the article. If you cannot state the page in sixty characters, the page does not know what it is about yet.
2. Write the description as an advertisement, not a summary. It has one job: earn the click over the four results around you.
3. Generate the block and paste it in, replacing any existing title, description or canonical — duplicates are a common reason Google ignores all of them.
4. Preview the share card with the Open Graph previewer before you post the link anywhere, because platforms cache the first version they see.
5. Verify on the live page with the SEO page checker. Plenty of CMS platforms quietly overwrite what you pasted.

That last step catches more problems than the other four combined. The tags you wrote and the tags being served are not always the same thing, and the only way to know is to look at what the server actually returns.

FAQ