01 / UX Research · Product Design
Meridian's logistics platform had grown from a simple tracking tool into a multi-module enterprise product serving operations teams, account managers, and executives simultaneously. Over four years of feature additions, the navigation had accumulated debt: three competing menu paradigms, inconsistent labeling, and a top-level structure that mirrored the engineering team's architecture rather than the mental models of the people using it.
A baseline usability audit revealed that 68% of users couldn't complete cross-module tasks without at least one navigation error. Onboarding time for new operations staff had grown to 11 days. Customer success was fielding the same "where is X?" tickets at scale.
The brief was deliberately open: reduce navigation complexity and cognitive load without disrupting power users who had learned to work around the existing structure. A redesign that broke muscle memory would create a different kind of churn.
The first six weeks were entirely research. I ran 22 contextual interviews across three user segments — operations coordinators, account managers, and executives — using a combination of think-aloud task walkthroughs and card sorting exercises. This surfaced a clear hierarchy: 80% of daily use concentrated in four core modules, while the remaining fourteen occupied the same visual weight in the existing nav.
I also ran a tree testing study with 140 participants to benchmark task completion rates and identify structural weak points. The data confirmed what interviews suggested: the problem wasn't labeling, it was hierarchy. Users had accurate mental models of what existed — they just couldn't predict where to find it.
I explored three structural directions: a role-based nav that adapted to user type, a frequency-weighted model that surfaced recent and frequent items, and a simplified flat hierarchy with a persistent command palette. Each was prototyped at medium fidelity and validated with a different set of users from the original research pool. The command palette approach consistently outperformed on task speed while the role-based model scored highest for first-use comprehension.
The final direction synthesized both: a simplified top-level structure with role-contextualized defaults, and a keyboard-accessible command palette for power users. Every design decision was tied back to a research insight documented in a shared Notion system I built to keep the engineering and product teams aligned throughout delivery.
−61%
Navigation error rate
6 weeks post-launch
4.2d
Average onboarding time
down from 11 days
+38%
Cross-module task
completion rate
The redesign shipped in three phases over twelve weeks, with a feature flag rollout that allowed us to A/B test at 20%/80% split before full deployment. Support tickets related to navigation decreased by 44% in the first month. The command palette, initially scoped as a stretch goal, became the most-used feature for the power-user segment within three weeks of launch.
The project also produced a reusable navigation component library and a decision framework for future information architecture decisions — both now part of Meridian's internal design system.