How many languages should an audio guide have?

Start with the local language, not the visitor list

The instinct is to treat an audio guide as a translation product and start with the languages of international visitors. That gets the order wrong.

Local-language listening is consistently the largest share of use at the venues we work with. Visitors who read the signage perfectly well still choose to listen, because audio is easier while walking, it works while your eyes stay on the object, and it carries tone that a label cannot. At Shantytown, which offers the guide free, English listening dominates daily use even though English is the local language.

That has a practical consequence. A strong local-language guide serves the largest group immediately, and it is also the source text every translation will be built from. Getting it right first makes every later language cheaper and better.

Use your own data, carefully

Most museums have some sense of where visitors come from, through ticketing, tour operator bookings, coach traffic, or visitor books. That data is worth using, but it answers a slightly different question than the one you are asking.

Nationality is not language. Visitors from Switzerland may want German, French, or Italian. Visitors travelling from one country may be citizens of another. And confident English speakers from anywhere may simply prefer English, which is why raw nationality counts tend to overstate translation demand.

The more useful signal arrives after launch. A web-based guide can show which languages visitors actually select, which turns the second round of translation into an evidence-based decision rather than a guess. Deciding language four to eight after six months of real usage is far more reliable than deciding all eight before opening.

• Ticketing and tour operator data show origin, which is a proxy rather than an answer
• Coach and cruise schedules are a strong signal where they apply
• Post-launch language selection data beats every pre-launch estimate
• Front of house staff usually know which languages they are asked for

A workable starting set

For most regional museums and heritage sites, three to five languages at launch is a sensible target: the local language, plus the two to four with the clearest evidence behind them.

That is enough to cover the majority of visitors who need translation, small enough to review properly, and small enough that the guide actually launches. A guide live in four languages is worth considerably more than a guide still in planning for twelve.

The exception is venues with a dominant single inbound market, such as a site on a fixed cruise or coach route. There, two languages done well can outperform six done thinly.

What each additional language actually costs

Museums usually budget for translation and stop there. The recurring costs are the ones that determine whether a multilingual guide survives.

Every language has to be reviewed by someone who speaks it, ideally someone who also knows the collection. Every future correction has to be applied in every language, or the versions drift apart until visitors in different languages are told different things. Every new exhibition raises the same question again.

This is why the delivery model matters as much as the translation budget. If adding or correcting a language means reprinting signage or rebooking a studio, languages get added once and then quietly rot. If it means updating content in a browser, the guide can keep pace with the museum.

• Translation is the one-off cost, review and maintenance are the recurring ones
• Corrections must propagate to every language or versions drift
• New exhibitions repeat the whole cycle
• Delivery model decides whether languages stay current or go stale

Let visitors find their language without work

Offering a language is not the same as a visitor finding it. A guide with eleven languages behind a menu the visitor has to hunt through will be used in one.

The entry point should do the work. Detecting the phone's language and opening the guide in it when that version exists removes the step entirely, and a visible language switcher handles the cases where the phone setting is not what the visitor wants.

Signage should also make the offer obvious in the languages themselves. A QR sign that says the guide is available in Mandarin, written in Mandarin, reaches people that an English-only sign never will.