The Smithsonian's National Museum of African American History and Culture built the Searchable Museum to bring its exhibitions to audiences who can't visit in person. The platform is ambitious: digitized collections, 3D models, videos, audio, all organized into an online experience meant to feel like walking through the galleries.
But the search feature wasn't working. Users typed in queries and got results that didn't match what they were looking for. They clicked on a result and landed in the middle of a page with no indication of where the matching content actually was. They tried to go back and lost their search entirely. The cognitive burden was high enough to cause real frustration.
We conducted five remote moderated sessions with educators, lifelong learners, and family researchers. Each participant searched for topics that mattered to them, and we watched what happened.
The problems showed up immediately. One participant searched for Harriet Tubman and got results about related topics, but nothing specifically about Harriet Tubman. The search was keyword-based and couldn't distinguish between content that mentioned a subject and content that was actually about that subject.
When participants clicked a result, the page loaded somewhere in the middle. They didn't know where to start reading or how to find the specific phrase they had searched for. Participants described feeling lost.
Without filters, some participants improvised. They used multi-phrase searches or tried advanced syntax to narrow results. These were workarounds for missing functionality.
And when participants opened an item from the Explore feature and then closed it, they didn't return to their search results. They had to start over. Small friction, but it accumulated.
We documented six findings, each with specific recommendations. Work with curators on taxonomy so search results connect to what users actually want: take users directly to the matching text and highlight it. Add lightweight filters. Let people return to their search. The fixes weren't dramatic, but they addressed real frustration.
The team implemented the priority changes. Search now includes filters, and results land where users expect them to.
