Choose one question the pilot must answer
A museum might want to test whether international visitors use translated audio, whether a highlights route improves independent visits, or whether people will pay for premium interpretation. Pick one primary question. It will shape the route, signage, pricing, and data you need.
Avoid setting vague goals such as 'try digital interpretation'. A measurable question creates a useful decision at the end: expand, change the visitor journey, add a language, adjust the offer, or stop.
Select 10 to 20 strong stops
Choose objects that already attract attention, need context, contain a human story, or prompt repeated visitor questions. Include an opening stop that explains the route and a closing stop that suggests what to do next.
Do not copy every wall label. Write for listening and use the object in front of the visitor. A practical track usually begins with a hook, adds the missing context, directs the visitor to notice something specific, and ends before attention drifts.
One welcome or orientation stop
A focused group of collection highlights
One or two stories that only audio can tell well
A closing prompt, donation message, or next-step invitation
A transcript for every spoken track
Design discovery before launch day
A guide cannot prove its value if visitors never see it. Plan the entrance poster, front-desk sentence, object labels, ticket or receipt message, and pre-visit information together. Decide whether a single route code or object-level codes make more sense for the space.
Run a staff walk-through with the actual signs in place. Ask someone unfamiliar with the project to enter as a visitor, find the guide, choose a language, and play a stop without coaching. Fix the points where they hesitate.
Measure a small set of useful signals
Track scans or starts, language choice, popular stops, drop-off, completed payments if the guide is paid, and common staff questions. Pair the numbers with short observation and visitor comments. Analytics can show what happened; conversations often explain why.
Set a review date before launch. Four to eight weeks is often enough for an operational pilot, although a highly seasonal venue may need to cover a particular visitor period. Record the decision and the evidence behind it so the next phase is deliberate.
Frequently asked questions
How many stops should an audio guide pilot include?
Ten to twenty focused stops is often enough to test discovery, content, languages, and visitor behaviour without creating a large production burden.
How long should a pilot run?
Four to eight weeks can reveal useful patterns, but seasonal venues should choose a period that represents the visitors they most need to understand.
Should a pilot be free or paid?
Match the pilot to the question. Use free access to test reach and engagement; use paid access when willingness to purchase is the main uncertainty.