When your website speaks the language but misses the person

There is a parlour trick that people who work with words and taxonomy develop over time. Scan a website for thirty seconds and you can tell, just from the labels, which market it was built for and whether it was actually localised for the one it's sitting in.

The table below does the heavy lifting

None of those differences is dramatic. None of them will cause a legal problem or a technical failure. But collectively, they create a feeling. A sense that the site was written for someone else and you just happen to be reading it.

That feeling costs you conversions.

This is the bit that trips up a lot of digital projects, particularly ones with international scope and a tight budget.

Translation takes content written in one language and converts it into another. If you take a website built for an American audience and translate it into French for a Canadian launch, you have translated it. The words are now French. The grammar is French. Base language: covered.

What you have not done is make it feel like it was written for someone in Quebec.

Localisation takes a dialect in one region and converts it into the appropriate dialect for another. Taking that same American English website and rewriting it into Québécois French, the dialect spoken in Quebec, is localisation. The vocabulary shifts, the idioms shift, and the cultural register shifts. The reader feels known, not merely served.

The distinction matters most in taxonomy and navigation. Labels, categories, filter names, and calls to action are all dialect-dependent. They are not just words. They are signals about whether you understand the person reading them.

When a site is translated rather than localised, the mismatch tends to show up earliest in the findability layer. Customers search using their own vocabulary, their dialect, their regional term for a thing. If the taxonomy was built in another dialect, the language of the labels will not match the language of the query.

That is a search failure caused by a vocabulary problem, and it is one of the most overlooked sources of missed revenue in e-commerce and digital services alike.

A customer in Manchester searching for "delivery options" should not have to work out that your site calls it "shipping." A customer in Edinburgh should not have to translate your category labels in their head before they can navigate. That cognitive load is invisible in analytics, but it is absolutely there, and it accumulates.

It is not just swapping words. Effective localisation of a website taxonomy requires:

The last point is where most projects stop short. Localisation delivered without testing is still a guess.

A translated site is functional, at best. Users can navigate it. They will not feel at home in it.

The sites that feel cold, over-formal, or slightly off are almost always sites where the localisation work was either skipped or conflated with translation. The language is technically correct. The register is wrong. And customers, who are not specialists in linguistics, will simply feel that the site was not made for them. They are right.

Getting the vocabulary right, at the taxonomy level, is one of the highest-leverage improvements a digital team can make. It does not require a redesign. It requires knowing which words your customers actually use, and making sure your site uses them too.

Want to know whether your site's taxonomy is speaking your customers' language? That is exactly what a Murmuration diagnostic looks at.