lev-os Information Architecture
Overview
Information Architecture (IA) is the structural design of shared information environments. It encompasses the art and science of organizing and labeling websites, intranets, online communities, and software to support usability and findability. The goal is to help users find information and complete tasks with minimal cognitive friction.
IA operates at the intersection of three critical elements (Rosenfeld & Morville's Information Ecology):
- Users: Their behaviors, needs, information-seeking patterns, and mental models
- Content: Document types, metadata, volume, structure, and governance
- Context: Business goals, constraints, culture, technology, politics, and resources
Good IA feels invisible—users navigate effortlessly without conscious thought. Bad IA creates confusion, frustration, and abandoned tasks. The discipline applies cognitive science, library science, and user research to make complex information spaces comprehensible.
When to Use
- Designing navigation for content-heavy sites
- Users can't find features or information
- Restructuring existing site architecture
- Building knowledge bases or documentation
- High bounce rates/low engagement suggest findability issues
The Process
Step 1: Research Phase - Understand Users and Content
Conduct stakeholder interviews to understand business goals and constraints. Perform content audit to inventory existing information assets. Analyze usage data (analytics, search logs) to identify popular content and failed searches. Interview users about their information needs and current pain points.
Example: E-commerce site audit reveals 1,500 products across 8 categories. Analytics show 40% of users use search (high = navigation failure). Search logs show users looking for "affordable laptops under $500" but taxonomy is brand-based, not price-based.
Step 2: Card Sorting - Map Mental Models
Use open card sorting (users create categories) to discover natural groupings. Conduct with 15-30 representative users. Analyze results for patterns—where do 70%+ of users agree? This reveals how users expect information to be organized, not how your company thinks about it.
Example: 20 users sort 50 features. 85% group "Settings," "Account," "Billing" together → Create "Account Management" section. Only 20% grouped "Reports" with "Analytics" → Keep separate despite internal team's assumption they belong together.
Step 3: Define Taxonomy and Hierarchy
Create organizational schemes (by topic, by task, by audience, chronological). Design navigation hierarchy—3 levels maximum for optimal findability. Broader categories at top, specific items nested. Apply the 7±2 rule: 5-9 items per category to prevent choice overload.
Example:
- Level 1: Products (top nav)
- Level 2: Electronics, Clothing, Home (category pages)
- Level 3: Laptops, Phones, Tablets (subcategories) Avoid going deeper—users get lost beyond 3 clicks.
Step 4: Create Navigation Systems
Design global navigation (primary site-wide menu), local navigation (contextual within section), and supplemental navigation (breadcrumbs, related links, sitemaps). Ensure consistent placement and behavior across the site.
Example: Global nav persists on all pages. Local nav appears in sidebar showing current section's subsections. Breadcrumbs show "Home > Electronics > Laptops > Dell XPS 13."
Step 5: Develop Labeling System
Use user language from research, not internal jargon. Keep labels concise (1-2 words ideal, 3 max). Make labels mutually exclusive—no overlap between categories. Test labels with 5-10 users: "If you needed [X], which category would you check?"
Example: Internal team calls it "Solutions Hub." Users don't search for "solutions." User research shows they say "help docs" or "guides." Rename to "Help Center" or "Guides."
Step 6: Design Search System
Implement site search for content-heavy sites. Use clear search scopes (search within current section vs. entire site). Display search results with relevant metadata (category, date, type). Provide filtering and faceted search for large result sets.
Example: Documentation site search shows article title, category tag, last updated date. Users can filter by "Getting Started," "API Reference," "Troubleshooting."
Step 7: Tree Testing - Validate Findability
Give users realistic tasks ("Where would you find your order history?"). Track task success rate, time, and path taken. Iterate if success rate <80% or users take wrong paths frequently. Tree testing validates IA before visual design investment.
Example: Tree test task: "Find instructions to reset your password." Success rate: 65% (too low). Most users checked "Account" but it was actually under "Help." Move password reset to "Account Settings" → Retest: 90% success.
Key IA Components
Organization Systems:
- Exact (alphabetical, chronological, geographical) - when users know specific name or date
- Ambiguous (by topic, by task, by audience) - when users browse for discovery
- Hybrid schemes for different content types within same site
Labeling Systems:
- Contextual links (hypertext links within content)
- Headings (page and section titles)
- Navigation labels (menu items)
- Index terms (keywords, tags for search and retrieval)
Navigation Systems:
- Global navigation (persistent site-wide)
- Local navigation (section-specific)
- Contextual navigation (related links, inline links)
- Supplemental navigation (sitemap, index, search)
Search Systems:
- Search algorithms (ranking, relevance)
- Query builders (advanced search interfaces)
- Retrieval results (presentation, sorting, filtering)
- Faceted navigation (progressive filtering)
Common Pitfalls
Organizational Silos: Mirroring company departments in navigation (Marketing, Sales, Engineering) instead of user tasks (Get Started, Pricing, Documentation). Users don't care about your org chart.
Deep Hierarchies: Going 5+ levels deep (Home > Products > Electronics > Computers > Laptops > Gaming > Budget > Dell). Users abandon before reaching content. Keep it 2-3 levels maximum.
Choice Overload: Presenting 15 top-level categories. Research shows 7±2 optimal. Beyond 9, decision paralysis sets in. Group related items into broader categories.
Inconsistent Labels: "Account" in header, "Profile" in footer, "My Settings" in mobile menu—all pointing to same place. Pick one term, use it everywhere.
Ambiguous Labels: "Resources" (what kind?), "Solutions" (to what problem?), "Products" (when you sell services). Be specific: "Documentation," "Case Studies," "SaaS Platform."
Ignoring Search Logs: Users tell you what they want via search queries. If 1,000 people search "cancel subscription," make it easy to find, don't bury it 4 levels deep.
No Escape Hatches: Users trapped in dead-end pages with no way to navigate elsewhere except back button. Always provide paths forward.
Anti-Patterns
- ❌ >7 top-level categories (choice overload)
- ❌ >3 levels deep (users get lost)
- ❌ Organizing by company structure vs. user mental models
- ❌ Vague labels ("Resources," "Solutions")
- ❌ Skipping user research—assuming you know how users think
- ❌ Inconsistent terminology across touchpoints
- ❌ No search for sites with >50 pages
Real-World Application
Situation: Internal knowledge base has 2,000 articles. Employees complain they "can't find anything." Support tickets increase because people can't self-serve.
IA Process:
- Audit: Catalog all 2,000 articles by topic, author, date, views.
- Analytics: Search logs show top 20 queries: "expense report," "PTO policy," "submit bug," etc.
- Card Sort: 25 employees sort 60 common articles. Patterns emerge: "HR & Benefits," "Engineering Tools," "Sales Resources," "Company Policies."
- Taxonomy: Create 5 top-level categories (not 15). Organize by task, not department.
- Labels: Use employee language from card sorts, not HR jargon. "Request Time Off" not "PTO Requisition."
- Search: Implement faceted search (filter by department, topic, format).
- Tree Test: 10 employees complete tasks ("Find maternity leave policy"). 90% success rate.
Outcome: Support tickets from "can't find X" drop 60%. Employee satisfaction with knowledge base increases from 2.3 to 4.1/5.
Complementary Frameworks
- User Research Methodologies (understand user needs)
- Usability Testing (validate navigation effectiveness)
- Content Strategy (govern what content exists)
- Interaction Design (design how users navigate)
Learning Resources
- "Information Architecture: For the Web and Beyond" by Rosenfeld, Morville & Arango (canonical IA text, 4th edition)
- OptimalWorkshop tools for card sorting, tree testing
- Nielsen Norman Group IA resources and case studies
Metadata
Domain: App Development, UX Design, Content Strategy Confidence: High - Field established since 1990s with mature methodologies Practitioner Weight: 10/10 (Rosenfeld & Morville are canonical practitioners) Execution Complexity: Medium - Clear methods, requires user research skills ROI Evidence: High - Improved findability directly impacts task completion and satisfaction
Related
- user-research-methodologies
- card-sorting
- tree-testing
- navigation-design
- taxonomy
- usability-testing
- content-strategy