Self-Service Kiosks
Client:
The Broad Museum
Role:
Year:
Explore the full story –
Challenge
Graffiti culture is broad and interconnected. Artists belong to movements, movements belong to decades, and techniques cut across all of it. Almost any structure you draw is defensible on paper. That is the problem. A kiosk gives you one standing visitor, no onboarding, and no way to ask a question. If the categories do not match how a visitor already thinks, they do not browse more slowly, they stop. And a designer cannot tell the difference between a structure that works and one that merely makes sense to the person who drew it, because the person who drew it already knows where everything is. So the real question was not how to organize the content. It was how to know whether my organization matched anyone else's.
Objective
Derive the structure from evidence instead of instinct, then test it before building on it. I ran an open card sort with five participants over moderated Zoom sessions using KardSort, asking them to group the exhibition's topics and name the groups themselves rather than choosing from categories I had already written. Then I turned the results into a proposed architecture and ran a separate usability test against it, giving participants real tasks and watching where they hesitated. Around that, the kiosk had its own conditions to design for: an attract screen for an idle machine, a sixty-second reset so one visitor's session does not become the next visitor's starting point, twelve languages, and image rights for artwork the museum does not own.
Results
The card sort produced consensus and disagreement, and the disagreement was the useful part. Some groupings were unanimous. Banksy and Basquiat landed together for four of five participants. But three cards split roughly evenly between Culture and History, and they split for a real reason: a mural from 1983 is genuinely both. The study recommended merging the two sections. I did not. Merging would have produced one oversized section and lost the distinction between what graffiti was and what it does, which is the distinction the exhibition is actually about. Instead I kept both and relabeled them so the boundary was legible from the outside, then confirmed the split held in usability testing. Following research is not the same as obeying it. The data told me where the ambiguity was. The decision about what to do with it was mine to make and defend. The result is a complete clickable prototype: one homepage into five sections and twenty-two topic pages, navigable by tile, by card, by section to section jump, and by previous and next. Underneath it sits a design system of twenty components and roughly 455 instances, with every color bound to a three-tier token set. Every topic carries sourced content, and a consolidated credits system tracks each image's attribution and license across public-domain, Creative Commons, and fair-use material.
