Case study
Rebuilding navigation around user goals instead of product silos
Cloudflare Zero Trust had grown into a powerful enterprise platform, but users consistently described the navigation as confusing, overlapping, and difficult to learn.
I led a large-scale information architecture initiative to redesign the navigation around mental models, workflows, and task clarity rather than internal product structures.
Impact snapshot
My impact
Validated the new structure at scale
Led multi-phase IA research with 393 survey responses, 273 tree test participants, qualitative walkthroughs, and statistically significant validation testing.
Reframed the navigation model
Moved the dashboard from product-centric organization toward observability, solution overview, use-case management, global configuration, and global settings.
Created scalable governance
Established principles for top-level additions, child pages, reusable systems, settings placement, and future feature expansion.
Introduced a decision framework
Created a structured flowchart to help teams decide where future capabilities should live within the IA.
Made language more literal
Reduced reliance on acronyms, internal product names, abstract concepts, and labels that required Cloudflare-specific context.
Expanded the UX strategy
Influenced longer-term thinking around contextual navigation, discoverability, search, education, and workflow connectivity.
The problem
The navigation reflected Cloudflare's org chart, not how users thought about their work
As Zero Trust expanded, the dashboard evolved organically around products, feature ownership, and technical terminology. Users repeatedly described the experience as fragmented, overlapping, difficult to scan, and difficult to learn. Or as one so aptly put it:
"The burden of navigation."
Another user said they would "end up with a dozen tabs open," sometimes with duplicates they did not realize were there.
The deeper issue was not just naming. Research showed that users approached the dashboard through workflows like troubleshooting a tunnel, securing an application, investigating traffic, managing devices, and configuring policies, not through Cloudflare's internal product boundaries.
The project quickly evolved from a navigation cleanup effort into a broader rethink of how the platform should organize itself around user intent.
The solution
A layered information architecture model
I introduced a five-layer IA framework that grouped the dashboard by how users engage with the platform.
Layer 1
Observability
What should I know about my environment?
Layer 2
Object ecosystem
What am I protecting and connecting?
Layer 3
Use-case management
How do I configure security workflows?
Layer 4
Global configurations
What reusable systems support multiple workflows?
Layer 5
Global settings
How do I manage the overall platform experience?
Research approach
Multi-phase research and testing
The work combined qualitative interviews, moderated walkthroughs, large-scale surveys, tree testing, and open-ended mental model analysis. It was intentionally iterative: I would test a structure quickly, study where users struggled, adjust the model, and test again.
The validation data was statistically significant, which gave the team more confidence that we were not just reacting to a handful of loud anecdotes. I used the research to evaluate findability, overlap, terminology clarity, user mental models, and conceptual understanding of reusable systems.
I brought those findings to product and engineering leaders to build the buy-in needed across 30+ stakeholders. That alignment helped move the work from a proposed IA cleanup to a shared direction for navigation strategy, governance, and future expansion.
Reflection
Navigation shapes how users understand complex systems
This project fundamentally changed how I think about enterprise UX systems. I left with three principles:
01
Labels should be literal, specific, and jargon-free
Navigation labels should clearly communicate what users will find or do. We need to avoid clever phrasing that may obscure meaning, and we also need to be careful with terms we assume are industry standard. According to our users, top-level items should feel mutually exclusive and collectively exhaustive. Specificity not only helps users scan, it is critical in building trust.
02
Users think about their tasks, not how our system works
Users approach the dashboard with goals, not technical models in mind. They look at navigation labels almost like a task list. When labeling navigation items, we need to prioritize task clarity over system accuracy. Imagine explaining the navigation structure to a user by saying, "Here, you will find X." Whatever X is should be the label. This means avoiding product names and acronyms when possible.
03
Navigation goes beyond the menu
Navigation happens everywhere. The side navigation menu cannot and should not be the only way users understand the system and make connections. Users have a deep desire for search, personalization like recents or favorites, clear in-workflow connections, and strong product content to reduce reliance on the side navigation alone.