What Google and the assistants see of your site

Most of the flaws you find online are not slips of attention. They are mistakes that nothing stopped from going out: text typed into a box and published unchecked — and that box is what the search engines and the assistants read.

The box and the page

On most sites, the title shown in search results comes from a field somebody fills in by hand, one day, without seeing the result. Nothing warns them if the text runs too long, if it will be cut off mid-word, or if it was left empty.

The outcome is not visible from the site. It is visible only where your customer looks for you.

The flaws that get through, every day

• The truncated title. Written too long, cut off by the engine mid-word. The first thing anyone reads of you is an unfinished sentence.
• The missing or duplicated description. The engine then makes one up itself, from whatever it finds. You no longer choose your own summary.
• Accents and capitals. Text entered without accents, or all in capitals, reads badly and quotes badly. In French, a missing accent sometimes changes the word.
• Empty modules. A carousel that no longer loads, a section left in placeholder text. The visitor sees a hole; the machine reads a blank.
• Forgotten files. A test folder, a backup page, a technical trace left reachable. What is reachable ends up indexed.

Why this now counts twice

These flaws already cost visits. Now they cost citations: an assistant asked to summarise your brand leans on the same text. A truncated title and an engine-invented description produce an invented answer as well.

A check that depends on no one

The length of a title, the presence of a description, spelling, accents, the pages that must not be public: all of it can be checked automatically, at every publication. What a machine verifies no longer depends on one person’s attention on a Friday evening.

It is also a question of alignment. A supplier who bills for the repair has no reason to prevent the breakdown. We would rather it did not happen.