MCN 2026

The best exhibit technology disappears. A visitor reaches for a screen, a button, and a story unfolds. That invisibility requires, in practice, a set of architectural decisions that many museum platforms never make — decisions that determine not just who can experience an exhibit today, but whether the system holding it together will still hold up after exhibits rotate, content changes, audience needs shift, staff turns over, and the vendors you launched with are no longer the vendors you have.

At the Obama Presidential Center, we treated accessibility and longevity as architecture problems. The result is a centralized exhibit platform powering the OPC's digital experiences. It spans cloud-based content authoring, a MongoDB-backed exhibit API, on-premises Docker services, and a fleet of devices — touchscreen kiosks and dedicated accessibility UXP devices — all drawing from the same content and event model. Every device, whatever its form, is a client of the same system. The visitor at the touchscreen and the visitor at the UXP station are not having parallel experiences maintained by separate codebases. They are having the same experience, delivered differently. Accessibility decisions in museums are often mistaken for content decisions — alt text, audio descriptions, alternative formats. These matter, but they are downstream of something more fundamental: whether the system itself was designed to serve more than one kind of visitor.

This didn't emerge from the architecture. It is the architecture.

Building the suite of services taught us that the decisions with the deepest impact on accessibility and longevity are rarely the visible ones — and that getting them right early is what makes everything else possible. The unglamorous choices about how content is shared, how devices communicate, and how much the running exhibit depends on live infrastructure turned out to be the choices that had the most impact.

This session is a case study in those decisions — how we made them, what they cost, what they produced, and what we would do differently. The OPC had dedicated engineering resources that many institutions cannot count on. We don't want to paper over that. But the principles that guided us are not proprietary, and the questions they answer — who can access this, and will it last — are ones every institution is already living with, whatever their stack.

I hope to leave attendees with a clearer sense of what trustworthy exhibit infrastructure can look like, and a vocabulary for advocating for it in their own contexts.