A prototype is a question made physical. It does not need to resemble the finished attraction in every respect; it needs to reproduce the conditions that matter to the question. Cardboard can reveal whether a control is discoverable. A full-scale taped footprint can reveal whether a group blocks a sightline. Neither proves that the final installation is safe or operationally ready.
Begin by naming uncertainty. “Test the room” is too broad to guide useful work. “Find out whether first-time visitors understand the shared objective without staff prompting” creates a clearer study. It identifies what to build, whom to invite, what to observe and what kind of design change might follow.
Choose fidelity by risk
Different questions require different kinds of realism. A narrative test may need convincing language and rough props. A reach study needs relevant dimensions and participant variation. A timing study needs a credible sequence, including waiting and reset. High visual polish can distract from those needs and make a weak concept feel harder to change.
Document what the prototype cannot tell you. A manually triggered light may test whether feedback is understood, but it says little about sensor reliability. A host quietly adjusting the pacing can conceal an operating problem. These limits do not invalidate the test; they define the evidence correctly and prevent the team from drawing unsupported conclusions.
Recruit beyond the project team
Designers know where controls are hidden and what the story is supposed to mean. Familiar colleagues may unconsciously provide the expected response. Include people who have not seen the concept, and recruit across relevant ages, access needs, language backgrounds and group types. Explain the activity clearly and obtain appropriate consent for participation and any recording.
Do not expose participants to unfinished hazards in pursuit of realism. Establish a controlled test environment, suitable supervision and clear stop conditions. Specialist review is necessary where a prototype introduces physical, electrical, motion or other risks. A rough build is not exempt from responsible planning simply because it is temporary.
Observe before explaining
Give participants the introduction the attraction actually intends to provide. Then watch what they do. Note the first object they approach, the point where they hesitate, the questions they ask and the cues they overlook. If a facilitator rescues the interaction, record the intervention rather than counting the task as independently understood.
Use neutral follow-up questions. “What did you think would happen?” is more informative than “Did you notice the confirmation light?” Ask participants to explain their role in their own words. Separate direct observations from interpretations: “three people waited beside the doorway” is evidence; “the entrance felt intimidating” is a possible explanation to investigate.
Rehearse the complete cycle
Once the core interaction is understandable, test its place in the larger journey. Include briefing, group entry, play, ending, exit and reset. Staff should practise ordinary assistance as well as interrupted sequences. Observe how one delayed group affects the next rather than resetting the test environment to ideal conditions after every run.
Agree on decision criteria before the session. The team might look for an independently understood first action, a recognisable ending or a reset process with no undocumented steps. These criteria are project hypotheses, not universal acceptance thresholds. A small formative study can reveal problems, but it cannot establish statistical certainty or substitute for formal inspections.
Close the loop with a decision record
For each finding, record the evidence, likely cause, proposed change, owner and next test. Keep unresolved issues visible. A successful session is not one in which participants praise the concept; it is one that helps the team decide what to retain, change or stop. Repeatedly observed confusion deserves more attention than a persuasive internal explanation.
The outline below translates these principles into a compact discussion aid. Edit the example, preview the result and copy or download a plain-text version. It stays in the page until you choose an export action and is not submitted to this website. Use it to organise the next conversation, not to certify capacity, accessibility, engineering or safety.
