The five things you pay for
Every audio guide budget breaks into scripting, translation, narration, signage, and delivery. Suppliers differ mainly in how many of those they include, which is why two quotes for the same guide can differ by a factor of five and both be honest.
Each is priced on a different unit, and knowing which unit tells you immediately how the number will move when the project grows.
Scripting: priced in staff hours, or per stop if outsourced
Translation: priced per word, per language
Narration: priced per recorded minute or per language, human or synthetic
Signage: priced per sign, plus design and a print run
Delivery: a subscription, a setup fee, a revenue share, or hardware
Sizing it before anyone quotes
You can estimate most of a guide before talking to a supplier, because the size of the content is something you control. A ninety second stop is roughly 220 to 250 spoken words. A twelve stop guide is therefore around 3,000 words of finished script, which is a short essay, not a book.
That number drives almost everything downstream. Translation is priced per word, so three additional languages means paying to translate 3,000 words three times. Narration is priced per minute, so twelve stops is about eighteen minutes per language. Signage is one sign per stop plus one at the entrance. Once you know the word count and the stop count, you can sanity check any quote you receive.
The cost that never appears in a quote is the one that usually dominates: the staff time to write those 3,000 words, check them with curators, and get them approved. Budget it as real money even though no invoice arrives, because it is the stage that decides whether the guide is any good and the stage most likely to delay the launch.
One-off costs and costs that come back
Scripting, translation, and narration are one-off. You pay again only when the content changes, which for a permanent collection might be every few years. Signage is close to one-off, and cheap enough per sign that reprinting a handful is not a budget event.
Delivery is the part that recurs, and its shape depends entirely on the route you choose. Handheld devices recur hardest: they need charging, cleaning, repair, replacement on a three to five year cycle, and desk staff time every single day the venue is open. That daily staff cost is usually larger than the hardware itself and almost never appears in the business case.
A browser-based guide moves the recurring cost into a subscription or a revenue share and removes the hardware line entirely. That is the honest comparison to make: not the headline price, but what each option costs in year three.
Where the money usually gets wasted
Translating before the source script is locked is the most expensive mistake available, because it means paying per word twice in every language you committed to. It is also the easiest to avoid: sign the English off properly, then start translation.
Recording all languages in one session before anyone has heard the finished result runs a close second. Voice the main language, listen to it in the actual gallery on a phone speaker, fix the pacing, and only then commission the rest.
Covering the entire collection is the third. A guide to forty objects costs three times a guide to twelve and is not three times as good, because visitors were never going to listen to forty. Start with the objects people already stop at.
Translating a script that is not yet locked
Voicing every language before hearing the first one in situ
Covering everything instead of the stops that matter
Buying hardware before testing whether visitors will use their own phones
Leaving signage design until after the audio, then paying for a rush print
Frequently asked questions
How much does it cost to make an audio guide?
It depends almost entirely on how many stops, how many languages, and whether you buy hardware. The controllable part is content size: a twelve stop guide is around 3,000 words of script and eighteen minutes of audio per language, and every per-word and per-minute quote you receive is priced against those numbers.
What is the most expensive part of making an audio guide?
Staff time on scripting, if you count it honestly, and hardware if you buy it. Scripting rarely appears in a budget because it is paid in hours rather than invoices, and handheld devices keep costing money every day through charging, cleaning, and desk staffing.
Does each extra language double the cost?
No. It adds a translation cost per word and a narration cost per minute, but the scripting, the signage, and the delivery are already paid for. One QR code serves every language you publish, so nothing physical changes when you add one.
Can a museum make an audio guide for free?
Not quite free, but close. Writing and recording in-house with a decent microphone costs staff time and a small amount of equipment. What you cannot avoid paying for is somewhere to host and serve the audio to visitors, and signage to point them at it.
Do you have to buy audio guide devices?
No. A guide delivered through a QR code opens in the browser on the visitor's own phone, which removes both the purchase and the daily handling cost. Devices only earn their place where visitor phone use genuinely is not viable.