Measure the activation in layers: who could encounter it, who noticed, who tried, who completed the intended action, what they understood, and whether the intended follow-up happened. Select one primary outcome before designing the interface. A high interaction count can indicate that a game is inviting; it cannot, by itself, show that visitors understood the brand or took a useful next step.
Why one engagement number hides the result
Michelis and Müller observed 660 passers-by across two weekend evenings at a set of gesture-based public displays in Berlin. They described an Audience Funnel: passing, viewing and reacting, subtle interaction, direct interaction, multiple interactions, and follow-up. The study offers a way to describe distinct behaviours; its particular transition rates belong to its place and installation. Michelis and Müller, International Journal of Human-Computer Interaction, 2011.
A second field study makes the distinction sharper. Parra and colleagues deployed two interactive displays for a CPR awareness campaign in a Brussels rail station. They observed more than 10,000 passers-by and more than 1,000 reactions, but reported that only 0.10% reached the final stage of visiting the campaign website. Those figures are from one public health campaign, not a benchmark for brand events. The useful lesson is that visible response and the campaign's chosen outcome can diverge. Parra, Klerkx and Duval, PerDis, 2014.
Together, the studies support reporting stages instead of presenting the busiest stage as the whole result. They do not establish that a particular interface caused sales, recall, or business growth.
Set a primary outcome and a short measurement chain
An activation may have different jobs. If the goal is product understanding, use a concise question after the experience or a short observational task: can the visitor explain what the product does? If the goal is a qualified conversation, define what counts as one and let staff record it consistently. If the goal is follow-up, offer a relevant next step and track its completion. A selfie, a button tap, or time spent may be useful diagnostic signals, but they are weak substitutes for those outcomes.
One workable measurement plan has five columns:
| Stage | Practical observation | What it cannot prove |
|---|---|---|
| Opportunity | People entering the visible approach area | That each person saw or understood it |
| Notice | People who pause, turn, or look toward the activation | That they know it is interactive |
| Try | Distinct people who begin a turn | That they complete or value the task |
| Complete | People who finish the intended interaction | That the brand message is understood |
| Follow-up | Relevant conversation, opt-in, or destination visit | That the activation alone caused a later sale |
Record the observation period, station uptime, staff assistance, and any queue. A broken station or a facilitator who starts every interaction changes what a completion rate means. Count people and sessions separately when one visitor may replay.
Connect the count to the experience's purpose
ISO's definition of usability in ISO 9241-11:2018 depends on specified users, goals, and context of use. This is a helpful discipline for event measurement: an installation can be usable for a brief, social discovery task yet unsuitable for an unassisted, detailed product comparison. State the context before interpreting the data.
If visitors move from the venue to a website, campaign links can identify traffic that used a tagged link in Google Analytics. That is evidence of that route, not a count of everyone who later searched for the brand or visited through another device. Staff notes and a short visitor question can explain why people did or did not proceed; avoid combining them into an invented conversion rate.
For an Archv interactive experience, the reporting brief should sit beside the creative brief. Choose the intended visitor action, instrument only the stages needed to diagnose it, and review the findings against the actual venue conditions. That yields a more useful answer than reporting that a screen was “engaging.”