Redesigning an Enterprise Platform for 10,000+ Users
How a systems-first approach to information architecture reduced task completion time by 40% across a complex enterprise platform serving four distinct user groups.
The existing platform was built incrementally over 5 years by multiple teams. Each team added features independently, resulting in 47 different views, inconsistent navigation patterns, and a 12-minute average task completion time. Users reported that they could not find what they needed without asking colleagues.
Started by mapping the entire tool ecosystem — 23 internal tools, 8 third-party integrations, and 4 distinct user roles. Conducted stakeholder interviews across departments to understand actual workflows versus assumed workflows. Built an information architecture model that consolidated 47 views into 12 modular workspaces organized by user intent, not organizational structure. Designed a progressive disclosure system that surfaces relevant tools based on user role and current context. Validated through 3 rounds of prototype testing with real users from each role.
40% reduction in average task completion time. 60% decrease in internal support tickets related to navigation. The modular architecture enabled 3 product teams to ship features independently without breaking the unified experience. Platform adoption increased from 64% to 91% within 6 months.
Context
Large organizations rarely have a single platform problem. They have an ecosystem problem. This enterprise SaaS company had grown through acquisition and internal development, resulting in a fragmented digital landscape that users navigated through tribal knowledge rather than intuitive design.
The platform served over 10,000 daily active users across four distinct roles: operations managers, analysts, field engineers, and executive stakeholders. Each group had different needs, different mental models, and different levels of technical fluency.
The Real Problem
The initial brief was "redesign the dashboard." After two weeks of ecosystem mapping, it became clear that the dashboard was a symptom, not the disease.
The real problems were:
- No unified information architecture — each product team had created their own navigation patterns
- Role-based needs were invisible — the same interface served analysts running complex queries and executives needing three key metrics
- Tool fragmentation — users routinely switched between 5-7 tools to complete a single workflow
- Documentation gap — critical institutional knowledge existed only in people's heads
Approach
Phase 1: Ecosystem Mapping
Before touching any interface, I mapped the entire digital ecosystem. This included:
- 23 internal tools and their relationships
- 8 third-party integrations and data flows
- 4 user roles with distinct workflow patterns
- 18 core task flows across all roles
This map became the foundation for every decision that followed.
Phase 2: Workflow Analysis
I shadowed users from each role for two weeks, documenting actual workflows versus intended workflows. The gap was significant — users had developed workarounds that added an average of 4 extra steps per task.
Phase 3: Architecture Design
Rather than redesigning screens, I designed a modular workspace system:
- 12 workspaces organized by user intent (not organizational structure)
- Progressive disclosure that adapts to user role and context
- Unified navigation that works consistently across all workspaces
- Cross-tool task flows that eliminate the need to switch between applications
Phase 4: Validation
Three rounds of prototype testing with real users from each role. Each round refined the architecture based on actual usage patterns, not assumptions.
Results
The impact was measurable within the first quarter:
- 40% reduction in average task completion time
- 60% decrease in navigation-related support tickets
- 91% platform adoption (up from 64%)
- 3 product teams now ship independently on the modular architecture
Key Insight
The most impactful design decision was not visual. It was structural. By organizing the platform around user intent rather than organizational hierarchy, we eliminated the cognitive overhead that was costing users 12 minutes per task.
Technology was not the constraint. Structure was.