IBM Symphony — Optimising a complex enterprise workflow platform for dramatic efficiency and satisfaction gains.

What this project achieved.

What was broken and why it was hard.

IBM Symphony is a complex enterprise workflow platform used by professionals across IBM. Over years of incremental development, the platform had accumulated inconsistent navigation patterns, overwhelming information density, and task flows that required too many steps for high-frequency operations. Experienced users had built workarounds. New users struggled to onboard.

• Inconsistent navigation patternsUsers couldn't predict where to find functionality across different areas of the platform
• Cognitive overloadToo much information presented simultaneously with no clear hierarchy or progressive disclosure
• High step-count for frequent tasksOperations that should take seconds required multiple confirmation screens and context switches
• No error feedbackUsers received no clear indication of what went wrong or how to fix it
• No progress visibilityMulti-step workflows gave no indication of completion status or next required actions

Understanding users before redesigning anything.

Workflow analysis and stakeholder interviews revealed that users weren't slow because they lacked skill — they were slow because the platform didn't match their mental models. I combined workflow observation, task analysis, and user interviews to build a clear picture of where the highest-value friction lived.

• Insight
  Mental model mismatch was the core problem. Users with years of experience still hesitated before common operations — not due to forgetting, but because the interface didn't map to how they thought about their work. Implication: IA restructuring around user mental models, not system architecture, was the highest-priority intervention.
• Insight
  80% of usage was concentrated in 20% of features. Workflow analysis showed the vast majority of daily tasks came from a small set of high-frequency operations. Implication: Optimising these core flows for speed and clarity would deliver the greatest efficiency gains.
• Insight
  Users needed less visible, not more visible information. Users consistently described feeling overwhelmed, yet also complained about not finding things — a classic progressive disclosure failure. Implication: Layer information by task context — show essentials first, details on demand.

Three principles that shaped every decision.

What changed and why.

The redesign focused on three structural improvements: unified navigation built around user task flows (not system modules), shortcut patterns for the top 20% of high-frequency operations, and real-time error feedback with progress indicators for multi-step workflows.

Measured outcomes.

The efficiency gains were concentrated in the high-frequency task set identified during research — exactly as the design strategy predicted. Users who had developed workarounds to cope with the previous design adopted the redesigned flows within their first session.

What I learned.

• The most valuable insight came from watching users work, not asking them what they wanted. Observation revealed the mental model mismatch that interviews alone never surfaced.
• Getting stakeholder consensus on what "efficiency" meant was harder than the design itself. Different teams had different success metrics. Alignment on measurement criteria was a prerequisite for the work.
• If I were doing this again, I'd involve engineers in the IA phase, not just the handoff phase. Technical constraints that emerged late in the process required IA adjustments that could have been avoided.