LEARN

Localisation Is Not Translation

Two arrays of funnels, one with wide shallow mouths and one with narrow deep mouths, each feeding its own dense core, representing content that must be rebuilt around local phrasing instead of carried over unchanged.
The cost of localization lies in reconstructing question sets; each side needs independent sets, as testing one side with the other's set yields meaningless results

IN ONE SENTENCE

Translated content usually earns citations on neither side — the phrasing, the priorities and the citable evidence are all different.

Translating content from one side to the other usually earns citations on neither: the phrasing differs, the priorities differ, and the citable evidence differs. Localisation has to change those three layers, not only the language.

OUR POSITION

Most of the cost of localisation sits in rebuilding the question set, not in translation. That is also why each side needs its own set — measuring one side with the other's questions produces numbers that mean nothing.

01

The three layers to change

Phrasing: the same decision question is asked differently, and a translated question is often not what locals actually type.

Priorities: for the same product, buyers in each market may ask different things first — compliance and integration on one side, cost and deployment speed on the other.

Evidence: the citable third-party sources differ, and so do the reference points buyers recognise. Cases and figures have to be recomposed against locally checkable standards.

02

Specific differences at the language layer

Chinese can carry high information density; English needs short sentences. Long translated sentences read stiffly in English, and stiffness costs professional credibility.

Terminology handling differs too: a term that is standard on one side may need a sentence of explanation on the other, and the reverse reads as padding.

03

One common technical error

⚠️ In hreflang across language versions, x-default must not have each locale pointing at the other — that declares two contradictory defaults. Point it at one fixed version, usually the primary market's.

With multi-domain deployments, each domain declares its own robots and sitemap. Do not share one static file with a hard-coded domain.

Sources

  1. [1]Optimizing your website for generative AI features on Google Search．Google Search Central．2026-05-15

Updated 2026-08-10