Meta Tag Fixes: Every Tag in the Head, Checked
Every web page carries a block of instructions in its <head> that visitors
never see: the title, the description, the canonical address, the robots rules, the viewport setting,
the language and the tags that social platforms read.
The Meta Tag Analyzer reads that block and checks each part: title length and uniqueness, description quality, whether the canonical is correct, Open Graph and X card completeness, robots directives and viewport configuration. This page treats the head as one system, explains what each finding means and how serious it is, and points you to the guide that fixes it.
What lives in the head, and who reads it
Meta tags are read by machines, not people. Search engines, browsers, screen readers and the crawlers behind link previews each pick out the tags they care about and ignore the rest. That is why one missing tag can break something in one place while everything else looks fine. The table below lists the tags the analyzer checks, who acts on each one and what goes wrong when it is missing or wrong.
| Tag | Main readers | What it controls | Typical fault |
|---|---|---|---|
<title> | Search engines, browser tabs, bookmarks | The usual source of the clickable headline in search results | Duplicated across a template, or the same brand name on every page with nothing else |
meta name="description" | Search engines | A candidate for the text under the headline | Missing, copied between pages, or a sentence that says nothing about the page |
link rel="canonical" | Search engines | Which address should be treated as the main version of the page | Pointing at the home page, at http, or at a different trailing-slash version |
meta name="robots" | Search engines | Whether the page may be indexed, followed, and how large a preview may be shown | A noindex left over from a staging
site |
meta name="viewport" | Mobile browsers | Whether the page is laid out for the screen width | Missing, so phones show a shrunken desktop layout |
html lang | Screen readers, browsers, translation tools | Pronunciation, hyphenation and translation prompts | Missing, or en-US on a British site |
og: tags | Link preview crawlers | The title, image and text of a shared link | No image, or a relative image address |
twitter: tags | X, and some chat apps | The card type and overrides for X | No twitter:card, so links get a small card |
None of these tags is a ranking lever on its own. Google has said for years that the meta keywords tag is ignored, and a well-written description does not push a page up the results. What meta tags do control is whether a page can be indexed at all, which address gets the credit, how the result and the shared link look, and whether the page works on a phone. Those are reasons enough to get them right.
The fix guides, grouped by job
Each guide covers one finding in full: what the analyzer flagged, why it matters, the correct markup, how to change it on common platforms and how to confirm the fix.
Tags that decide how search engines treat the page
Fixing title length and uniqueness
covers writing one distinct title per page, putting the words people search for near the start, and deciding
when a brand suffix earns its space. Fixing
meta description quality covers descriptions that summarise the page honestly, mention what the searcher
is looking for and are not copied from page to page. Fixing
missing or broken canonicals covers the self-referencing canonical and the common mistakes: http
instead of https, inner pages pointing at the home page, and slash or parameter mismatches. Fixing
robots meta directives covers accidental noindex, the preview directives such as max-image-preview:large,
and when noimageindex does more harm than good.
Tags that shape shared links
Fixing Open Graph tag completeness covers
the core set of og: tags, plus og:site_name and og:locale, and
an image that works on the main platforms. Fixing
X (Twitter) card markup covers the twitter:card types and when explicit twitter:
tags are worth adding on top of Open Graph.
Tags that affect every visitor
Fixing a missing viewport tag covers
the one-line tag that makes a responsive layout work on phones, and why blocking zoom is an accessibility
problem. Fixing a missing or wrong lang
attribute covers choosing between en, en-GB and en-US, marking
up passages in other languages, and how lang differs from hreflang.
Where to change the tags on your platform
Most meta tag problems are made by a template, a theme or a plugin, so the fix happens there too.
Meta tags in WordPress covers
Yoast, Rank Math and SEOPress, which setting wins when the post and the site disagree, and bulk editing.
Meta tags in Shopify covers the
head of theme.liquid, product and collection overrides, and apps that print duplicate tags.
Meta tags in Next.js covers the
Metadata API, generateMetadata for dynamic routes, generated share images and why tags must
be in the server-rendered HTML.
Which findings to fix first
The analyzer reports everything it finds, but the findings are not equal. Some stop a page appearing in search at all; others only make a result or a shared link a little less appealing. Work from the top of this list down.
- An unintended
noindex. A page carryingnoindexwill drop out of Google once it is recrawled, however good it is. It is often copied across from a staging site or switched on by a plugin setting, and it can affect a whole site at once. - A canonical pointing somewhere else. If every product page names the home page as its canonical, search engines may treat the products as duplicates. A canonical is a hint rather than a command, but a wrong one is a strong hint in the wrong direction.
- A missing viewport tag. Without it, phones lay the page out at a desktop width and shrink it, and Google indexes the mobile version of pages, so the mobile experience is the one that counts.
- Duplicate or empty titles. Pages with identical titles are hard to tell apart in results and in browser tabs, and they make it harder for search engines to see what each page is for.
- Missing or duplicate descriptions. Google writes its own snippet when there is no useful description, which is not always a disaster, but a repeated sentence across hundreds of pages wastes the one chance you have to suggest the text.
- Incomplete Open Graph and X card tags. These do not affect search, but every link shared without an image or proper title looks broken.
- A missing or wrong
lang. Quick to fix, and it matters to screen reader users, who may hear English read with the wrong pronunciation rules.
noindex or changing canonicals across a site,
check the tag was not put there on purpose. Thank-you pages, internal search results, filtered listings
and print versions are often excluded deliberately.Fix the template, not the page
Meta tag problems usually repeat. A site with 400 pages rarely has 400 separate problems; it has three or four template rules producing the same fault on every page that uses them. The quickest route through an analyzer report is to group findings by page type, such as blog posts, products, categories and service pages, and fix the rule that writes the tag for each type.
Where the tags come from
- The theme or layout file prints a base set of tags on every page, often including a default title and description.
- An SEO plugin or app adds its own set, usually with per-page overrides and site-wide patterns such as page title, separator, site name.
- Page-level fields filled in by editors override the pattern for that page only.
- Scripts in the browser can change tags after the page loads. Search engines may or may not pick those changes up, and link preview crawlers never do.
When two of these layers both print tags, you get two titles, two descriptions or two canonicals, and the reader picks one unpredictably. The analyzer flags duplicates, and the fix is to switch one layer off rather than to edit both.
Check what the server sends
Browser developer tools show the page after scripts have run, which can hide problems. To see exactly what a crawler receives, use View Source, or request the page from the command line:
curl -s https://example.com/services/ | grep -iE "<title|name=\"description|rel=\"canonical|name=\"robots|name=\"viewport"If a tag appears in the developer tools but not in that output, it is being added by JavaScript, and it should be moved into the server-rendered HTML.
X-Robots-Tag HTTP header, which
never appears in the HTML. If a page will not index and the head looks clean, check the response headers
with curl -I.Tags that contradict each other
Some of the most confusing findings come from tags that are each valid on their own but disagree with each other. Search engines then have to guess which signal you meant, and they do not always guess the way you would like.
| Conflict | What happens | How to resolve it |
|---|---|---|
noindex plus a canonical pointing to another page | Two opposite messages: drop this page, and treat that one as the main version. Google has advised against combining them. | Decide
which you mean. For duplicates, use the canonical alone; for pages that should not appear, use noindex
alone. |
Canonical and og:url name different addresses | Search engines credit one address while shares are counted against another. | Generate both from the same value in the template. |
Robots meta says one thing, X-Robots-Tag says another | The most restrictive rule is generally the one applied. | Set robots rules in one place only. |
| Canonical points to a page that redirects or returns an error | The hint points nowhere useful, so it is likely to be ignored. | Point the canonical at the final, working address. |
lang="en-US" on content written for the UK | Screen readers and translation tools apply the wrong locale. | Use en-GB, or plain en if you do not
target a region. |
Two <title> elements | Browsers use the first; other readers may not. | Remove the duplicate at its source, usually the theme or a second plugin. |
Canonical choices also interact with hreflang on multi-language sites, where each language version should name itself as canonical. That topic has its own hub: Hreflang fixes.
A sensible order of work
- Run the Meta Tag Analyzer on one page of each type: the home page, a service or category page, a product or article, and a contact page.
- Sort the findings by page type and look for the ones that repeat. Those point to a template rule.
- Deal with anything that blocks indexing first: stray
noindextags, wrong canonicals and missing viewport tags. - Remove duplicate tags by switching off one of the layers that prints them, then set sensible site-wide patterns for titles and descriptions.
- Rewrite titles and descriptions by hand for the pages that matter most, such as the home page, top services and best-selling products.
- Complete the sharing tags and set
langcorrectly in the layout. - Run the analyzer again on the same pages, then on a wider sample, to confirm the template changes worked everywhere.
This hub is about the head as a whole. Two neighbouring hubs go deeper on one part each: SERP preview fixes looks at how the finished result appears in Google, and Open Graph fixes looks at building and validating the sharing tags. For the full reference, read the Meta Analyzer Guide or open the example report.
1. Do meta tags still matter for SEO?
Some matter a great deal and some not at all. The robots tag and canonical decide whether a page can be indexed and which address receives the credit, so errors there can remove pages from search. The title is usually the source of the headline in results. The description can shape the snippet. The meta keywords tag is ignored by Google. So meta tags rarely lift rankings on their own, but mistakes in them can cost you visibility and clicks.
2. How long should a title tag be?
There is no fixed limit in the HTML standard, and Google does not cut titles at a set number of characters. It shortens them to fit the width available, which depends on the device and on the letters used. Around 50 to 60 characters usually fits without being cut on desktop. More important than the exact count is that each title is unique, describes the page accurately and puts the most useful words first, so a shortened version still makes sense.
3. Why does Google show a different description from the one I wrote?
Google treats the meta description as a suggestion. It often builds the snippet from text on the page that matches the particular search more closely, so one page can show different snippets for different searches. Your description is more likely to be used when it accurately summarises the page, mentions what people are searching for and is not repeated on other pages. A missing or generic description almost guarantees that Google writes its own.
4. What is a canonical tag and does every page need one?
A canonical tag is a line in the head that names the preferred address for a page. It helps search engines choose between near-identical versions, such as the same page with and without tracking parameters, or with and without a trailing slash. It is good practice for every indexable page to carry a canonical that points to its own clean address. It is a strong hint rather than an instruction, so search engines can choose a different canonical if other signals disagree.
5. How do I find out if a page has an accidental noindex?
Run the page through the Meta Tag Analyzer, or open View Source and search for noindex. Also check the HTTP response headers, because a noindex can be sent in an X-Robots-Tag header that never appears in the HTML. In Google Search Console, the URL Inspection tool reports whether indexing is allowed, and the Pages report lists addresses excluded by a noindex tag. On WordPress, check that the setting that discourages search engines from indexing the site is switched off.
6. Should I use en-GB or en in the lang attribute?
Use en-GB if your content is written in British English for a UK audience, and plain en if it is written for English speakers generally. Both are valid. The value helps screen readers choose pronunciation and helps browsers with spelling checks, hyphenation and translation offers. It does not tell Google which country to show the page in; for sites with separate versions for different countries or languages, that job belongs to hreflang annotations, which are a separate set of tags.
7. Why do I have two title tags on my page?
Usually because two parts of the site are both printing one. A theme may write a title into the head, and an SEO plugin or app writes another, or a page builder adds its own. Browsers use the first title they find, but other readers may take the second, so the results are unpredictable. Find which layer should be in charge, normally the SEO plugin, and turn off title output in the other one rather than editing both copies.
The Meta Tag Analyzer reads the head of your page and checks the title, description, canonical, robots directives, viewport, Open Graph and X card tags, then lists each finding with the guide that fixes it.
Run the Meta Tag Analyzer →