Meta Tag Analyzer: How It Works
A meta tag analyser fetches a live page and reports what search engines and social platforms actually receive — which is frequently not what the CMS preview promised. It is the fastest way to catch the small head-section errors that quietly suppress a page.
What gets checked
| Element | Healthy | Flagged when |
|---|---|---|
| Title | Present, unique, 50–60 chars | Missing, duplicated, over 60, or stuffed |
| Description | Present, 150–160 chars | Missing, truncated, or duplicated site-wide |
| Canonical | Absolute, self-referencing, 200 status | Relative, missing, or pointing at a redirect |
| Robots | index, follow | noindex on a page you want indexed |
| H1 | Exactly one | None, or several competing |
| Open Graph | title, description, image, url, type | Missing image, or an image under 1200×630 |
| Viewport | Present | Missing — breaks mobile rendering |
| Language | lang on the html element | Missing or wrong |
The errors worth panicking about
Most findings are minor. Three are not:
noindexin production. Almost always left over from a staging environment. It removes the page from search entirely, and it is invisible to anyone looking at the rendered page.- A canonical pointing elsewhere. If page A canonicalises to page B, you are instructing search engines to drop A. A single templating error can do this across an entire section.
- A
Disallow: /in robots.txt. Not a meta tag, but the same category of catastrophe, and worth checking at the same time.
Each of these can remove a working page from search overnight and produce a traffic drop that looks like a penalty.
Rendered versus raw
A meaningful complication: many sites set meta tags with JavaScript. A tool fetching raw HTML sees the server response, which may contain placeholder or missing tags, while Google's renderer executes the JavaScript and sees the final values. If an analyser reports missing tags on a page you know sets them client-side, verify with Search Console's URL inspection, which shows the rendered HTML. Server-rendered tags remain the safer approach — they work for every crawler, not just the ones that execute JavaScript.
Beyond a single page
Analysing one page catches page-level errors. Site-wide problems need site-wide checks:
- Duplicate titles and descriptions across many URLs — the most common finding in any large-site audit.
- Canonical chains, where A points to B which points to C.
- Mixed signals — a page that is in the sitemap, canonicalises to itself, and carries
noindex. - Parameterised URLs generating hundreds of near-identical pages.
What a clean report does not mean
Passing every check means your technical signals are consistent, nothing more. It says nothing about whether the content answers the query, whether the page is faster than competitors, whether anyone links to it, or whether it deserves to rank. Meta tags are the hygiene layer: getting them wrong hurts, getting them right removes an obstacle rather than creating an advantage.