Creating Effective Data-Driven Information Architecture
Financial Advisor Portal IA Redesign
Project Overview
The Challenge
A major financial services company's advisor portal had poor findability, with staff struggling to locate critical information quickly. Content was organised around internal business structure rather than user needs and tasks.
The Impact
Card sorting session revealing how financial advisors naturally group content and tasks
Defining the Problem
How might we redesign the portal's information architecture to match financial advisors' mental models and task flows, enabling quick access to critical client information?
Initial Constraints
- Complex content: Hundreds of financial products, regulations, and tools needed logical organisation
- Diverse users: New advisors and experienced staff had different information needs and familiarity levels
- Business constraints: Existing content management systems limited radical structural changes
- Time pressure: Advisors needed quick access to information during client meetings
Research Process
Current State Analysis
Conducted comprehensive audit of existing content and user analytics to understand current pain points and usage patterns before designing new structures.
- Content inventory and classification of 200+ pages
- Analytics analysis revealing high bounce rates and search failures
- Stakeholder interviews to understand business priorities
- User journey mapping to identify critical information touchpoints
Open Card Sorting
Facilitated card sorting sessions with 24 financial advisors to understand how they naturally categorise and group financial content and tools.
IA Design & Tree Testing
Developed multiple IA proposals based on card sorting insights, then tested navigation effectiveness using tree testing methodology:
- Scenario-based tasks: Realistic client meeting scenarios requiring information finding
- Success metrics: Time to find, success rate, and confidence ratings
- Iterative refinement: Three rounds of testing with progressively improved structures
- A/B comparison: Current vs. proposed IA performance comparison
Validation & First-Click Testing
Conducted first-click testing with interactive prototypes to validate the final IA design and identify any remaining navigation issues.
- High-fidelity prototype testing with realistic content
- Task-based scenarios reflecting real-world advisor needs
- Heat map analysis showing navigation patterns and pain points
Key Findings
Task-Based Mental Models
Financial advisors organised information around client meeting stages and conversation topics, not company departmental structures.
Context-Dependent Navigation
The same content needed to be findable through multiple pathways depending on the advisor's current task or client situation.
Experience Level Differences
New advisors needed more guided pathways and explanatory content, while experienced staff preferred direct access to specific tools.
Search vs. Browse Preferences
Different advisors had strong preferences for either search-first or browse-first approaches, requiring both pathways to be optimised.
Results & Impact
Significant Findability Improvement
Tree testing showed 46% improvement in task success rates, with average time to find information reduced by 38%. First-click testing confirmed these improvements held up in realistic interaction scenarios.
User Satisfaction & Adoption
"Finally, a system that thinks like we do. I can find what I need for client meetings without hunting through departments that don't make sense to me." — Senior Financial Advisor
Post-launch surveys showed 89% of advisors found the new structure more intuitive, with significant improvements in portal usage frequency and session duration.
Business Impact
The improved findability translated to measurable business outcomes, with advisors able to access client-relevant information 40% faster during meetings.
- Task-based navigation structure aligned with client meeting flow
- Dual-pathway design accommodating different user preferences
- Progressive disclosure for different experience levels
- Cross-linking strategy enabling multiple pathways to same content
Key Learnings
Mental Models Trump Organisational Structure
The most significant improvements came from aligning the IA with users' task-based mental models rather than internal company organisation. This required challenging stakeholder assumptions about "logical" content groupings.
Iterative Testing Reveals Hidden Issues
Each round of tree testing revealed subtly different issues that wouldn't have been apparent from card sorting alone. The iterative approach was essential for achieving the final performance improvements.
Multiple Pathways Increase Satisfaction
Rather than forcing all users through one "optimal" path, providing multiple ways to reach the same content accommodated different working styles and significantly improved overall satisfaction.