Theo Langlois

My whole approach treats these as plumbing rather than marketing, because that is what they are and how they behave. The title and description are read at request time and shown in a result listing. Both are truncated by the engine at a width rather than a character count, so the number of characters that fits depends on what is in them. A title containing a wide character occupies more horizontal space than one containing narrow ones, which is a detail worth knowing before optimising to a target length.

The canonical tag states which URL should be indexed for a piece of content. It is a hint, not a command, and it is ignored when it points somewhere that redirects. A canonical that contradicts the redirect chain is the most common form of the signal split. The robots meta tag controls indexing for one page.

Its interaction with the robots file is settled and not intuitive: a page disallowed in the file cannot be crawled, so its meta tag never gets read, and cannot therefore be allowed by it. That asymmetry causes a great deal of confusion. The sitemap is a list of URLs, and what matters in it is accuracy. Sitemaps carrying redirects, non-canonical URLs or pages that are blocked waste the budget they exist to spend.

I finish on the order these get processed in, because that determines which instruction wins when two of them disagree. None of these tags is a ranking factor on its own. They affect whether a page can be crawled, whether it can be indexed, and which of several addresses is treated as the one. That is the whole of their job.