The Title Suffix That Broke 207 Titles

01Every title was under the limit, until it was not

A directory site I run has 265 listing pages, each built from one template, and the template was careful. It capped every title at 54 characters, comfortably under the 60 that Bing flags as too long. Bing flagged them anyway. When I measured the titles as they actually appeared in the page, 207 of the 265 were over 60. The template was right about its own work and wrong about the page. Templates count the characters they write. Search engines count the characters you publish.

The second layer

The cause was a separate layer. The site layout appends a brand suffix to every page title, a separator and the site name, 15 characters in all. The listing template never saw it, because it was added after the template finished, by code written months earlier for a different purpose. Each layer was correct on its own. Together they produced titles of up to 69 characters, cut off in results and flagged in the webmaster report. Nobody had ever looked at the final string, because nobody owned it.

02One owner for the whole title

Let the page decide

The fix gave the listing template ownership of the complete title, suffix included. It now builds the full string, measures it, and keeps the brand suffix only when the whole thing fits in 60 characters. Fifty eight titles kept their suffix unchanged. The rest dropped it. And because the old cap had been trimming useful detail to make room for a suffix it could not see, 47 titles got that detail back, in this case the commission rate, which is exactly what a person comparing listings wants to read. A brand name in every title is worth less than the one fact that makes someone click.

Consistency without repetition

Dropping the brand from a title does not hide who you are. The site name still appears in the page, in the structured data and in the address, and search results show the site name separately. What matters for recognition is that the name is the same everywhere it appears, the point of entity hygiene, not that it is crammed into every title. The same site still has blog titles that carry the brand twice, up to 132 characters long. They are on the list, waiting for a decision, because they are older pages that already rank.

03Test the output, not the parts

Render, then measure

The general lesson is about where you test. Any check that runs inside one layer of a page can be undone by the next layer. The title cap lived in the template; the suffix lived in the layout. A description check that lives in an article generator can be undone by a site wide default. So the length check now runs on the rendered page: build it, read the final title and description out of the HTML, measure them, and fail on anything outside 10 to 60 characters for titles or 50 to 160 for descriptions. A check that cannot see the final page is a check on your intentions, not your output. It is the same reason I check the origin and not only the public copy, after the pages that only existed at the edge.

One more for the table

This joins the list I started in write down every search error once, along with a smaller fix shipped the same day: a breadcrumb link inside the structured data that pointed at an address which redirects, so every listing reported a redirect in its own markup. Both were template problems. Both were invisible in a browser. The bugs that reach every page are always in the code that every page shares. When a warning appears on hundreds of pages at once, look at the layers they have in common before you look at any single page.

The same audit found 39 listing descriptions between 161 and 192 characters and 46 other pages up to 265. Those are next, with the owner's go, because they sit on pages that already rank.

Tomorrow: a guide to how long a meta description should really be.