Finding the right media: Redesigning Search at Imago
Year:
2023
Brand:
Imago
Duration:
6 months
Industry:
Media
When we talked to users, the pattern was consistent. They wanted the freedom to explore loosely when they had a vague idea in mind, and the ability to zero in immediately when they knew exactly what they needed. One user put it simply:
"I end up scrolling past hundreds of photos because typing what I actually mean into the search bar never gets me the right result."
Julia, Magazine editor
Before
After
The Solution
We started by going back to basics: auditing every category in the system and mapping out a clear, logical information architecture. This gave us a foundation that matched how users actually grouped and thought about images, rather than how the backend happened to store them.
Insights
Search quality is a design problem, not just an engineering one. Sitting with engineers early, rather than handing off a spec, meant the neural clustering actually solved the user problem, not just a technical one.
Speed and flexibility aren't opposites. Users didn't want to choose between broad and precise search. They wanted both, available at the moment they needed it.
Typos and imprecise language are real user behavior, not edge cases. Designing (and training the model) for how people actually type, not how they're supposed to type, closed a real gap in the experience.
