Title tag too long or too short

Search results are laid out in pixels, not characters. A title of capital letters and wide glyphs truncates sooner than one of narrow lowercase letters, so any character count is an approximation of a pixel width.

The width also differs between devices and result types, and it changes when Google alters its result layout — which it does.

Google's own guidance describes what makes a good title link in terms of being descriptive and concise, and explicitly does not publish a character limit.

So the useful discipline is not counting. It is ordering: assume the tail may be invisible and put nothing essential there.

A truncated title is not a ranking problem. The full text is read; only the display is cut.

The cost is the click. A result reading `Acme Sports Equipment Store — Official Site — Blue Run…` has spent its visible width on words identical across every page of the site.

The brand suffix is the usual culprit. It is a convention, not a requirement, and on a deep page it consumes the most valuable part of the line to say something the domain already says.

Google also rewrites titles it judges unhelpful — over-long, keyword-stuffed, or repeated across many pages — and a rewritten title is a signal that the original was weak rather than a malfunction.

The opposite failure is quieter and more common. `Products`, `Blog`, `Page 2`, a single product code: a title that is present, unique, and identifies nothing to someone reading a list of results.

On catalogues it usually comes from a template emitting one field where two are needed — the product name without the category, or the category without the qualifier that distinguishes it.

A short title is also a wasted opportunity in a different sense: the title is the strongest single descriptor a page gives about its own subject, and leaving it at one word discards that.

The test is whether someone who sees only the title, with no URL and no description, could tell what the page contains.

`document.title` in the console returns the string. Counting its length is a rough proxy; looking at it in a real result is better.

Search for the page directly — `site:example.com/page` — and see what is actually rendered. That is the only measurement that accounts for pixel width.

Across a site, look at the distribution rather than individual pages: how many titles fall far outside the useful range, grouped by template. One template is one fix.

Compare against what Google shows. A page whose displayed title differs from the one you wrote has been rewritten, and that is worth reading as feedback.

Most specific part first. `Blue running shoes, sizes 36-47 — Acme` beats `Acme — the best shop for all your sporting needs`.

Add at most one qualifier. Loading every attribute into the title makes it unique, unreadable, and truncated before the useful part appears.

Drop or shorten the brand suffix on deep pages where the space is better spent. Keep it on the home page and top-level sections, where the brand is the subject.

Fix it in the template, and make sure the fallback also produces something identifying. A template whose default title is the site name reintroduces duplication wherever a field is blank.

Where two pages resist being told apart in a title, treat that as information rather than as a formatting problem: it usually means they should be one page.

`seo.title_length` reports titles far outside the range that displays usefully, as information rather than as a failure — a long title is truncated, not rejected.

`seo.title` reports the absent and empty ones, and `seo.duplicate_titles` reports the groups sharing a value, which is the adjacent problem a length report often surfaces.

The report gives length and uniqueness, both of which are facts. Whether a title reads well is a judgement, and the report does not score it.