Which languages should you translate into?

Why nationality data misleads

Regional visitor numbers describe the country, not your venue. A museum on a cruise itinerary, a museum next to a university, and a museum an hour inland see completely different mixes, and the national figures average all of them away.

Nationality is also not the same as language need. Plenty of international visitors read English comfortably and will use the English guide by choice. Others share a language with a nationality the statistics count separately. The question is not which passports arrive, it is which visitors are currently getting less out of the visit than they could.

The people who know this are on your front desk. They can tell you which groups arrive with a translating companion, which ones photograph panels to translate later, and which ones leave early. That is better evidence than a tourism board table, and it is free.

The local language is usually the biggest win

Across the New Zealand museums using Journey Pal, between 50% and 60% of guide use is in English, the local language. That is our own usage data rather than an industry estimate, and it consistently surprises museums that think of a guide as primarily a translation product.

The reason is that audio does something for local visitors that has nothing to do with language. It is easier while walking, it keeps your eyes on the object instead of on a panel, and it carries tone that print cannot. So the first and largest gain is often publishing well in the language most of your visitors already speak, then translating outward from there.

That also produces a better source text. A guide written properly in one language and then translated is stronger in every language than one where the source was rushed to get to translation faster.

Start narrow, then let the data decide

Three or four languages is a good first list: your local language plus the two or three with the clearest evidence behind them. It is achievable, reviewable, and it produces something better than a longer list done thinly.

Then look at what visitors actually select. A digital guide can tell you which languages are being opened, which is the only evidence that reflects your venue rather than the country. Museums are regularly surprised: a language they were confident about goes unused, and one they had not considered turns up steadily.

Adding a language later is a content change rather than a rebuild, so there is no penalty for starting small. The signage does not change, the codes do not change, and the guide does not need republishing anywhere.

• Local language first, done properly

• Two or three more with real evidence behind them

• Publish, then read which languages visitors actually open

• Add from the data rather than from the original guess

The constraint is reviewers, not translators

You can buy a translation in almost any language. What you cannot always find is someone who will check it before it goes live, and publishing a language you cannot review is a bet that nobody who speaks it will notice a mistake.

So treat the reviewer as the gate. If you can name the person or organisation who will check a language, it belongs on the list. If you cannot, it goes on the waiting list until you can, however good the visitor numbers look.

Indigenous and minority languages sit slightly outside this framework. The reason to publish te reo Māori is not visitor volume, and the review process is a relationship rather than a purchase. That deserves its own decision, made with the right people, rather than a slot in a ranked list.