Design the experience around the social unit that arrives, not only the person whose hand reaches the control. Decide whether companions can choose together, watch and advise, take turns, or each receive a separate result. A screen that works for one tester in a studio may leave the other two people in a group with nothing to do.
Assign useful roles before adding more controls
In a group interaction, not everyone must touch the screen. One person may operate, another compare options, and a third interpret the result aloud. If the experience is a racing challenge, spectators can see the goal and progress while the player controls the turn. If it is a product comparison, everyone should be able to understand which option is being examined without crowding the input area. These are design examples, not measured results from an Archv project.
Research on public displays shows why single-user assumptions need testing. Azad and colleagues observed group behaviour around mall directories, ticketing kiosks, and photo kiosks, then studied two individuals or two pairs using a simulated large vertical display. Their work identified changing physical territories around a shared screen and social conventions that affect parallel use. It does not establish that every event should support simultaneous input; it shows that the space around a display is part of the interface. Azad et al., Designing Interactive Systems, 2012; Waterloo research record.
Brief the roles: who starts, who can change a selection, whose result is shown, and what happens if two people act at once. A shared interface needs a visible ownership cue, a way to undo, and a clear end to one group's turn. Adding two touch points without deciding these rules can produce accidental conflict rather than collaboration.
Expect interruptions and conversation
Feuchtner, Walter, and Müller studied a gesture-controlled multiuser public-display game with 704 users. People often interrupted to socialize, and the tested manual pause techniques were explored but rarely used for their intended pause function. The authors argue that interruptions should be accommodated in the application design rather than assuming players will operate a conventional pause control. Their study is one game in a public setting, not a universal event rule. Feuchtner et al., MobileHCI, 2016.
This suggests a practical test: can a visitor look away to speak to a friend and still understand where the experience is when they return? Consider short self-contained steps, persistent cues about progress, and a clear “try again” route. Do not force a companion to read a dense instruction page while the active participant advances without them. If a session includes a private choice, give the participant control over what appears on the public screen.
Test with real groups, not isolated volunteers
Invite pairs and small groups who have not seen the demo. Observe how they approach, who takes control, whether others contribute, and how a second group waits or joins. Ask each person what they understood afterward. If only the operator can explain the result, the spectator path may be decorative. If everyone tries to control one task, the roles may be unclear.
This article addresses what happens within and between groups; station throughput and queue size are separate questions covered in Archv's capacity guide. The venue still needs a safe layout review, especially where bystanders gather around a display.
Archv's Formula 1 experience case includes racing-simulator concepts; it does not publish evidence of how groups used them. For a new installation, group behaviour should be observed in a prototype and written into the creative and operational handoff.