Rank markets before translating anything
Not every market deserves the same treatment, and pretending otherwise spreads budget and reviewer time thin. Most global teams work in tiers. Tier one markets get transcreated headlines, locally cast imagery and full legal review. Tier two get adapted copy and imagery chosen from the global set. Tier three get an accurate translation of the master with the right local legal lines. Rank markets by revenue at stake, regulatory exposure and cultural distance from the source market, and revisit the ranking each year. A market should move up a tier when it grows, and it has to move up when a local rule, such as a mandatory language requirement, makes adaptation unavoidable.
Glossaries and claims belong to the locale
The glossary is where localisation succeeds or fails at volume. Each locale needs a termbase: product names that stay in English, terms that must be rendered one specific way, and words that are banned because they mean something else locally or carry legal weight. Claims need the same discipline. A claim approved in one country can be unlawful in the next, so keep a claims matrix that states, per market, which claims are approved, which need a qualifier and which are off limits. Language and locale are not the same thing: Brazilian and European Portuguese, or Mexican and Castilian Spanish, need separate glossaries even when one translator could write both.
Native reviewers with a deadline and a veto
Machine output in thirty languages is only as good as the people who check it. Give each market a named reviewer who is a native speaker, knows the category and has the authority to reject. Give them a fixed review window and a short checklist: glossary terms used, claims matching the matrix, tone right for the brand in that language, no text broken by the layout. Feed their corrections back into the glossary so the same mistake is not fixed by hand in every campaign. Watch for reviewer fatigue. Someone asked to review forty assets a day starts approving on a skim, so plan the review load as carefully as the production load.
Sequencing a master into local versions
The order of work matters more than the tools. Lock the master first: final source copy, approved claims, approved imagery. Next, test layouts with longer and shorter text, because German and Finnish usually take more space than English while Chinese and Japanese take less, and layouts must flex before real translations arrive. Then localise copy, voice and on-screen text, rebuild layouts per script, including mirrored layouts for right-to-left languages, and send the set to market review and legal. In the studio, each market's glossary and approved claims sit in the brand hub and are applied at generation, so copy in 30 languages arrives with the terminology already right. Any later change to the master should reopen every derivative, not only the one someone remembers.
Updated 25 September 2026 · General information, not legal advice. Rules change, so check the current text with your legal team before relying on it.