Initial commit

This commit is contained in:
Zhongwei Li
2025-11-30 08:38:26 +08:00
commit 41d9f6b189
304 changed files with 98322 additions and 0 deletions

View File

@@ -0,0 +1,211 @@
{
"criteria": [
{
"name": "User Research Validation",
"description": "IA decisions based on user research (card sorting, tree testing, interviews), not assumptions or opinions.",
"scale": {
"1": "No user research. IA based on stakeholder opinions, org chart, or assumptions about users.",
"3": "Some user research (5-10 participants) but limited scope or informal methods.",
"5": "Rigorous user research: card sorting (15-30 users), tree testing (20-50 users), validated with real users before implementation."
}
},
{
"name": "Mental Model Alignment",
"description": "Navigation matches user mental models and vocabulary, not internal company jargon or org structure.",
"scale": {
"1": "Labels use internal jargon, org chart structure, or vague terms ('Solutions', 'Services'). Users confused.",
"3": "Mix of user-friendly and internal terms. Some alignment with user mental models.",
"5": "Labels match user vocabulary from research. Task-based or user-goal oriented. Clear mental model alignment."
}
},
{
"name": "Taxonomy Quality (MECE)",
"description": "Categories are mutually exclusive, collectively exhaustive. Clear, non-overlapping structure.",
"scale": {
"1": "Categories overlap significantly. Gaps in coverage. Items fit multiple places or nowhere.",
"3": "Mostly MECE but some overlaps or gaps. Most items have clear home.",
"5": "Perfect MECE: every item has exactly one primary home, no gaps, no overlaps. Faceted where appropriate."
}
},
{
"name": "Navigation Depth & Breadth",
"description": "Optimal depth (3-4 levels) and breadth (5-9 items per level). Not too flat or too deep.",
"scale": {
"1": "Too flat (>12 top-level items, overwhelming) OR too deep (>5 levels, users lost). Poor balance.",
"3": "Acceptable depth/breadth but some levels have <3 or >12 items. Room for optimization.",
"5": "Optimal: 3-4 levels deep, 5-9 items per level. Balanced hierarchy, manageable choices at each level."
}
},
{
"name": "Information Scent",
"description": "Clear labels, trigger words, descriptive breadcrumbs guide users. Strong predictive cues.",
"scale": {
"1": "Vague labels ('Resources', 'Other'), no breadcrumbs, users guess. Weak scent.",
"3": "Some clear labels but inconsistent. Breadcrumbs present but could be more descriptive.",
"5": "Strong scent: specific labels, user trigger words, descriptive breadcrumbs, preview text. Users click confidently."
}
},
{
"name": "Findability Metrics",
"description": "Measurable findability: time to find, success rate, search vs browse balance.",
"scale": {
"1": "No metrics tracked. Unknown if users can find content.",
"3": "Some metrics tracked (analytics, basic usability) but incomplete or informal.",
"5": "Comprehensive metrics: tree test ≥70% success, time to find <30 sec (simple) / <2 min (complex), 40-60% search/browse balance."
}
},
{
"name": "Tree Testing Validation",
"description": "Navigation structure validated with tree testing before implementation.",
"scale": {
"1": "No tree testing. Built navigation without validation.",
"3": "Informal tree testing (5-10 users) or partial testing (only some tasks).",
"5": "Rigorous tree testing: 20-50 users, 8-12 tasks, ≥70% success rate, <1.5× directness, iterated based on results."
}
},
{
"name": "Multiple Access Paths",
"description": "Users can find content via navigation, search, filters, tags, related links. Supports different strategies.",
"scale": {
"1": "Only one access method (browse-only or search-only). Single path to content.",
"3": "Two access methods (browse + search) but limited. No filters, tags, or related links.",
"5": "Multiple paths: browse, search, faceted filters, tags, breadcrumbs, related links. Supports searchers and browsers."
}
},
{
"name": "Scalability & Evolution",
"description": "Structure designed for growth. Handles 10× content without breaking. Governance plan exists.",
"scale": {
"1": "Structure works for current size but breaks at scale. No governance. No plan for growth.",
"3": "Some consideration for scale. Ad-hoc governance. Will need significant rework as content grows.",
"5": "Designed for scale: faceted for large sets, versioned taxonomy, governance framework (roles, processes, metrics), evolution plan."
}
},
{
"name": "Progressive Disclosure",
"description": "Starts simple, reveals complexity on-demand. Hub-and-spoke, collapsed sections, tiered navigation.",
"scale": {
"1": "Everything visible at once (overwhelming) OR everything hidden (users can't discover).",
"3": "Some progressive disclosure but inconsistent or incomplete.",
"5": "Well-designed progressive disclosure: hub-and-spoke for guides, collapsed sections for long nav, contextual secondary nav."
}
}
],
"guidance_by_type": {
"E-commerce Navigation": {
"target_score": 4.2,
"key_criteria": ["Taxonomy Quality (MECE)", "Multiple Access Paths", "Findability Metrics"],
"common_pitfalls": ["Too many top-level categories", "No faceted filters", "Search vs browse not balanced"],
"specific_guidance": "Use faceted navigation (Category × Price × Brand × Rating). Card sort with 20+ customers to understand grouping preferences. Tree test with 'Find product under $X in Y category' tasks. Track search vs browse ratio (target 40-60% each). Provide multiple paths: browse categories, search, filter, sort."
},
"Documentation IA": {
"target_score": 4.0,
"key_criteria": ["Mental Model Alignment", "Progressive Disclosure", "Information Scent"],
"common_pitfalls": ["Feature-based (not task-based)", "Too deep nesting", "No 'Getting Started' hub"],
"specific_guidance": "Structure by user tasks ('How do I...?') not features. Hub-and-spoke: 'Getting Started' hub with 5-7 clear entry points. 3-4 levels max. Breadcrumbs show path. Related links for discovery. Search with code snippet results. Tree test with 'Find how to do X' tasks (≥80% success)."
},
"Content-Heavy Website": {
"target_score": 3.8,
"key_criteria": ["User Research Validation", "Taxonomy Quality (MECE)", "Scalability & Evolution"],
"common_pitfalls": ["No content audit", "Overlapping categories", "No governance"],
"specific_guidance": "Start with content audit: inventory all content, identify duplicates/gaps. Open card sort (20-30 users) to discover categories. MECE taxonomy design. Tree test (30-50 users, 10-12 tasks). Plan for 3× content growth. Establish governance: taxonomy owner, quarterly reviews, tag moderation."
},
"SaaS Product Navigation": {
"target_score": 4.0,
"key_criteria": ["Mental Model Alignment", "Navigation Depth & Breadth", "Information Scent"],
"common_pitfalls": ["Internal jargon", "Org chart structure", "'Modules' label"],
"specific_guidance": "Card sort with active users (not stakeholders). Rename internal terms to user language: 'Widgets' → 'Reports', 'Modules' → 'Features'. Task-based labels: 'Create Report', 'Share Dashboard'. Flat structure (2-3 levels). Contextual help. Tree test with 'Where would you do X?' tasks. Iterate until ≥75% success."
},
"Intranet / Knowledge Base": {
"target_score": 3.5,
"key_criteria": ["Findability Metrics", "Multiple Access Paths", "Scalability & Evolution"],
"common_pitfalls": ["Outdated content", "Weak search", "No metadata schema"],
"specific_guidance": "Content audit: remove 40%+ outdated, merge duplicates. Closed card sort with employees to validate categories. Hybrid approach: browse by department + robust search + best bets for common queries. Metadata schema: author, last updated, audience, content type, tags. Governance: content owners responsible for freshness."
}
},
"guidance_by_complexity": {
"Simple (<100 items, clear structure)": {
"target_score": 3.5,
"focus_areas": ["Mental Model Alignment", "Navigation Depth & Breadth", "Information Scent"],
"acceptable_shortcuts": ["Informal card sort (10 users)", "Simple tree test (10 users, 5 tasks)", "Basic analytics"],
"specific_guidance": "Flat structure (1-2 levels). Simple categories. Informal user research acceptable (10 users). Basic tree test (5 tasks, ≥70% success). Monitor with analytics (bounce rate, time on site)."
},
"Standard (100-1000 items, moderate complexity)": {
"target_score": 4.0,
"focus_areas": ["User Research Validation", "Taxonomy Quality (MECE)", "Tree Testing Validation", "Findability Metrics"],
"acceptable_shortcuts": ["Card sort with 15 users (not 30)", "Tree test with 20 users (not 50)"],
"specific_guidance": "Card sort (15-20 users). MECE taxonomy (3-4 levels, 5-9 items per level). Tree test (20-30 users, 8-10 tasks, ≥70% success, ≤1.5× directness). Multiple access paths (browse + search). Track findability metrics (time to find, success rate, search/browse ratio)."
},
"Complex (>1000 items, high complexity)": {
"target_score": 4.5,
"focus_areas": ["All criteria", "Rigorous validation", "Comprehensive metrics"],
"acceptable_shortcuts": ["None - full rigor required"],
"specific_guidance": "Full content audit. Card sort (30 users). Faceted navigation (orthogonal facets). MECE taxonomy + controlled vocabulary. Tree test (50 users, 12 tasks, ≥75% success). Multiple access paths (browse, search, filters, tags, related links). Comprehensive metrics dashboard. Governance framework (taxonomy owner, quarterly reviews, versioning, deprecation process)."
}
},
"common_failure_modes": [
{
"name": "Org Chart Navigation",
"symptom": "Navigation structured by company departments (Sales, Marketing, Engineering) not user tasks or mental models.",
"detection": "Tree test shows low success (<50%). Users say 'I don't know what department handles this.' Card sort reveals users group differently than org structure.",
"fix": "Restructure by user tasks or goals. Card sort with users to discover natural groupings. Rename categories to user-facing labels: 'For Developers' not 'Engineering Department'."
},
{
"name": "No User Research",
"symptom": "IA designed by stakeholders or designers based on opinions, not user data. 'I think users will understand...'",
"detection": "No card sort, tree test, or user interviews conducted. Decisions justified by 'makes sense to me' not data.",
"fix": "Conduct card sorting (20-30 users) to understand mental models. Tree test (20-50 users) to validate proposed structure before building. Iterate based on results."
},
{
"name": "Too Deep Hierarchy",
"symptom": "Navigation 5-6+ levels deep. Users need 6+ clicks to reach content. Users get lost, high abandonment.",
"detection": "Tree test shows low directness (>2.5×). Analytics shows high bounce rate on intermediate pages. Users complain 'can't find anything'.",
"fix": "Flatten structure: 3-4 levels max. Add faceted filters or search to shortcut hierarchy. Use progressive disclosure (hub-and-spoke) instead of deep nesting."
},
{
"name": "Vague Labels",
"symptom": "Labels like 'Resources', 'Solutions', 'Services' — users don't know what's inside. Weak information scent.",
"detection": "Tree test shows users clicking multiple categories before finding content (low directness). Card sort shows users rename these categories to more specific terms.",
"fix": "Use specific, descriptive labels. 'Code Samples' not 'Resources'. 'API Documentation' not 'Solutions'. Test labels with tree testing. Match user vocabulary from card sort."
},
{
"name": "Non-MECE Taxonomy",
"symptom": "Items fit multiple categories OR don't fit anywhere. Overlap and gaps.",
"detection": "Content audit shows many items marked 'belongs in multiple places'. Users confused where to look. Card sort shows weak clustering (<50% agreement).",
"fix": "Redesign taxonomy to be mutually exclusive (no overlap) and collectively exhaustive (no gaps). Use faceted classification for items with multiple attributes. One primary location + facets/tags for cross-category access."
},
{
"name": "No Tree Testing",
"symptom": "Built navigation structure without validating users can find content. Launch reveals users struggle.",
"detection": "Live site shows low task success, high bounce rate, users complain 'can't find X'. No pre-launch validation done.",
"fix": "Always tree test before building. 20-50 users, 8-12 tasks. Target ≥70% success, ≤1.5× directness. Iterate until targets met. Saves rework post-launch."
},
{
"name": "Single Access Path",
"symptom": "Only browse (no search) OR only search (no browse). Users with different strategies struggle.",
"detection": "Analytics shows 90%+ use search (navigation broken) OR 90%+ browse (search broken). Users complain 'I can't browse' or 'search doesn't work'.",
"fix": "Provide multiple paths: browse (navigation), search, faceted filters, tags, breadcrumbs, related links. Support both 'searchers' (know what they want) and 'browsers' (exploring)."
},
{
"name": "Doesn't Scale",
"symptom": "Structure works for 100 items but breaks at 1000. Categories become massive, overwhelming.",
"detection": "One category has 60%+ of items. Users complain 'too much content'. Search becomes only viable path.",
"fix": "Design for scale upfront. Use faceted navigation for large sets. Split large categories into subcategories. Plan for 10× growth. Establish governance to evolve taxonomy as content grows."
},
{
"name": "No Governance",
"symptom": "Taxonomy degrades over time. Content in 'Other', empty categories, inconsistent tagging.",
"detection": ">10% content in 'Other' or 'Uncategorized'. Empty categories exist. User-generated tags have synonyms, typos, noise.",
"fix": "Establish governance framework: taxonomy owner, content owners, quarterly reviews. Monitor metrics (% in 'Other', empty categories, tag quality). Process for adding/removing categories, merging tags."
},
{
"name": "No Progressive Disclosure",
"symptom": "Everything visible at once (overwhelming mega-menu) OR everything hidden (users can't discover).",
"detection": "Users complain 'too much information' (overwhelming) OR 'I didn't know that existed' (hidden). Low engagement with deep content.",
"fix": "Progressive disclosure: hub-and-spoke for guides (overview → details), collapsed sections in navigation (expand on click), tiered navigation (primary always visible, secondary contextual)."
}
],
"minimum_standard": 3.5,
"target_score": 4.0,
"excellence_threshold": 4.5
}

View File

@@ -0,0 +1,494 @@
# Information Architecture: Advanced Methodology
This document covers advanced techniques for card sorting analysis, taxonomy design, navigation optimization, and findability improvement.
## Table of Contents
1. [Card Sorting Analysis](#1-card-sorting-analysis)
2. [Taxonomy Design Principles](#2-taxonomy-design-principles)
3. [Navigation Depth & Breadth Optimization](#3-navigation-depth--breadth-optimization)
4. [Information Scent & Findability](#4-information-scent--findability)
5. [Advanced Topics](#5-advanced-topics)
---
## 1. Card Sorting Analysis
### Analyzing Card Sort Results
**Goal**: Extract meaningful patterns from user groupings
### Similarity Matrix
**What it is**: Shows how often users grouped two cards together
**How to calculate**:
- For each pair of cards, count how many users put them in the same group
- Express as percentage: (# users who grouped together) / (total users)
**Example**:
| | Sign Up | First Login | Quick Start | Reports | Dashboards |
|--|---------|-------------|-------------|---------|------------|
| Sign Up | - | 85% | 90% | 15% | 10% |
| First Login | 85% | - | 88% | 12% | 8% |
| Quick Start | 90% | 88% | - | 10% | 12% |
| Reports | 15% | 12% | 10% | - | 75% |
| Dashboards | 10% | 8% | 12% | 75% | - |
**Interpretation**:
- **Strong clustering** (>70%): "Sign Up", "First Login", "Quick Start" belong together → "Getting Started" category
- **Strong clustering** (75%): "Reports" and "Dashboards" belong together → "Analytics" category
- **Weak links** (<20%): "Getting Started" and "Analytics" are distinct categories
### Dendrogram (Hierarchical Clustering)
**What it is**: Tree diagram showing hierarchical relationships
**How to create**:
1. Start with each card as its own cluster
2. Iteratively merge closest clusters (highest similarity)
3. Continue until all cards in one cluster
**Interpreting dendrograms**:
- **Short branches**: High agreement (merge early)
- **Long branches**: Low agreement (merge late)
- **Clusters**: Cut tree at appropriate height to identify categories
**Example**:
```
All Cards
|
____________________+_____________________
| |
Getting Started Features
| |
____+____ _____+_____
| | | |
Sign Up First Login Analytics Settings
|
____+____
| |
Reports Dashboards
```
**Insight**: Users see clear distinction between "Getting Started" (onboarding tasks) and "Features" (ongoing use).
### Agreement Score (Consensus)
**What it is**: How much users agree on groupings
**Calculation methods**:
1. **Category agreement**: % of users who created similar category
- Example: 18/20 users (90%) created "Getting Started" category
2. **Pairwise agreement**: Average similarity across all card pairs
- Formula: Sum(all pairwise similarities) / Number of pairs
- High score (>70%) = strong consensus
- Low score (<50%) = weak consensus, need refinement
**When consensus is low**:
- Cards may be ambiguous (clarify labels)
- Users have different mental models (consider multiple navigation paths)
- Category is too broad (split into subcategories)
### Outlier Cards
**What they are**: Cards that don't fit anywhere consistently
**How to identify**: Low similarity with all other cards (<30% with any card)
**Common reasons**:
- Card label is unclear → Rewrite card
- Content doesn't belong in product → Remove
- Content is unique → Create standalone category or utility link
**Example**: "Billing" card — 15 users put it in "Settings", 3 in "Account", 2 didn't categorize it
- **Action**: Clarify if "Billing" is settings (configuration) or account (transactions)
---
## 2. Taxonomy Design Principles
### Mutually Exclusive, Collectively Exhaustive (MECE)
**Principle**: Categories don't overlap AND cover all content
**Mutually exclusive**: Each item belongs to exactly ONE category
- **Bad**: "Products" and "Best Sellers" (best sellers are also products — overlap)
- **Good**: "Products" (all) and "Featured" (separate facet or tag)
**Collectively exhaustive**: Every item has a category
- **Bad**: Categories: "Electronics", "Clothing" — but you also sell "Books" (gap)
- **Good**: Add "Books" OR create "Other" catch-all
**Testing MECE**:
1. List all content items
2. Try to categorize each
3. If item fits >1 category → not mutually exclusive
4. If item fits 0 categories → not collectively exhaustive
### Polyhierarchy vs. Faceted Classification
**Polyhierarchy**: Item can live in multiple places in hierarchy
- **Example**: "iPhone case" could be in:
- Electronics > Accessories > Phone Accessories
- Gifts > Under $50 > Tech Gifts
- **Pro**: Matches multiple user mental models
- **Con**: Confusing (where is "canonical" location?), hard to maintain
**Faceted classification**: Item has ONE location, multiple orthogonal attributes
- **Example**: "iPhone case" is in Electronics (primary category)
- Facet 1: Category = Electronics
- Facet 2: Price = Under $50
- Facet 3: Use Case = Gifts
- **Pro**: Clear, flexible filtering, scalable
- **Con**: Requires good facet design
**When to use each**:
- **Polyhierarchy**: Small content sets (<500 items), clear user need for multiple paths
- **Faceted**: Large content sets (>500 items), many attributes, users need flexible filtering
### Controlled Vocabulary vs. Folksonomy
**Controlled vocabulary**: Preset tags, curated by admins
- **Example**: "Authentication", "API", "Database" (exact tags, no variations)
- **Pro**: Consistency, findability, no duplication ("Auth" vs "Authentication")
- **Con**: Requires maintenance, may miss user terminology
**Folksonomy**: User-generated tags, anyone can create
- **Example**: Users tag articles with whatever terms they want
- **Pro**: Emergent, captures user language, low maintenance
- **Con**: Inconsistent, duplicates, noise ("Auth", "Authentication", "auth", "Authn")
**Hybrid approach** (recommended):
- Controlled vocabulary for core categories and facets
- Folksonomy for supplementary tags (with moderation)
- Periodically review folksonomy tags → promote common ones to controlled vocabulary
**Tag moderation**:
- Merge synonyms: "Auth" → "Authentication"
- Remove noise: "asdf", "test"
- Suggest tags: When user types "auth", suggest "Authentication"
### Category Size & Balance
**Guideline**: Aim for balanced category sizes (no one category dominates)
**Red flags**:
- **One huge category**: "Other" with 60% of items → need better taxonomy
- **Many tiny categories**: 20 categories, each with 2-5 items → over-categorization, consolidate
- **Unbalanced tree**: One branch 5 levels deep, others 2 levels → inconsistent complexity
**Target distribution**:
- Top-level categories: 5-9 categories
- Each category: Roughly equal # of items (within 2× of each other)
- If one category much larger: Split into subcategories
**Example**: E-commerce with 1000 products
- **Bad**: Electronics (600), Clothing (300), Books (80), Other (20)
- **Good**: Electronics (250), Clothing (250), Books (250), Home & Garden (250)
### Taxonomy Evolution
**Principle**: Taxonomies grow and change — design for evolution
**Strategies**:
1. **Leave room for growth**: Don't create 10 top-level categories if you'll need 15 next year
2. **Use "Other" temporarily**: New category emerging but not big enough yet? Use "Other" until critical mass
3. **Versioning**: Date taxonomy versions, track changes over time
4. **Deprecation**: Don't delete categories immediately — mark "deprecated", redirect users, then remove after transition period
**Example**: Software product adding ML features
- **Today**: 20 ML-related articles scattered across "Advanced", "API", "Tutorials"
- **Transition**: Create "Machine Learning" subcategory under "Advanced"
- **Future**: 100 ML articles → Promote "Machine Learning" to top-level category
---
## 3. Navigation Depth & Breadth Optimization
### Hick's Law & Choice Overload
**Hick's Law**: Decision time increases logarithmically with number of choices
**Formula**: Time = a + b × log₂(n + 1)
- More choices → longer time to decide
**Implications for IA**:
- **5-9 items per level**: Sweet spot (Miller's "7±2")
- **>12 items**: Users feel overwhelmed, scan inefficiently
- **<3 items**: Feels unnecessarily nested
**Example**:
- 100 items, flat (1 level, 100 choices): Overwhelming
- 100 items, 2 levels (10 × 10): Manageable
- 100 items, 4 levels (3 × 3 × 3 × 4): Too many clicks
**Optimal for 100 items**: 3 levels (5 × 5 × 4) or (7 × 7 × 2)
### The "3-Click Rule" Myth
**Myth**: Users abandon if content requires >3 clicks
**Reality**: Users tolerate clicks if:
1. **Progress is clear**: Breadcrumbs, page titles show "getting closer"
2. **Information scent is strong**: Each click brings them closer to goal (see Section 4)
3. **No dead ends**: Every click leads somewhere useful
**Research** (UIE study): Users successfully completed tasks requiring 5-12 clicks when navigation was clear
**Guideline**: Minimize clicks, but prioritize clarity over absolute number
- **Good**: 5 clear, purposeful clicks
- **Bad**: 2 clicks but confusing labels, users backtrack
### Breadth-First vs. Depth-First Navigation
**Breadth-first** (shallow, many top-level options):
- **Structure**: 10-15 top-level categories, 2-3 levels deep
- **Best for**: Browsing, exploration, users know general area but not exact item
- **Example**: News sites, e-commerce homepages
**Depth-first** (narrow, few top-level but deep):
- **Structure**: 3-5 top-level categories, 4-6 levels deep
- **Best for**: Specific lookup, expert users, hierarchical domains
- **Example**: Technical documentation, academic libraries
**Hybrid** (recommended for most):
- **Structure**: 5-7 top-level categories, 3-4 levels deep
- **Supplement with**: Search, filters, related links to "shortcut" across hierarchy
### Progressive Disclosure
**Principle**: Start simple, reveal complexity on-demand
**Techniques**:
1. **Hub-and-spoke**: Overview page → Detailed pages
- Hub: "Getting Started" with 5 clear entry points
- Spokes: Detailed guides linked from hub
2. **Accordion/Collapse**: Hide detail until user expands
- Navigation: Show categories, hide subcategories until expanded
- Content: Show summary, expand for full text
3. **Tiered navigation**: Primary nav (always visible) + secondary nav (contextual)
- Primary: "Products", "Support", "About"
- Secondary (when in "Products"): "Electronics", "Clothing", "Books"
4. **"More..." links**: Show top N items, hide rest until "Show more" clicked
- Navigation: Top 5 categories visible, "+3 more" link expands
**Anti-pattern**: Mega-menus showing everything at once (overwhelming)
---
## 4. Information Scent & Findability
### Information Scent
**Definition**: Cues that indicate whether a path will lead to desired information
**Strong scent**: Clear labels, descriptive headings, users click confidently
**Weak scent**: Vague labels, users guess, backtrack often
**Example**:
- **Weak scent**: "Solutions" → What's in there? (generic)
- **Strong scent**: "Developer API Documentation" → Clear what's inside
**Optimizing information scent**:
1. **Specific labels** (not generic):
- Bad: "Resources" → Too vague
- Good: "Code Samples", "Video Tutorials", "White Papers" → Specific
2. **Trigger words** (match user vocabulary):
- Card sort reveals users say "How do I..." → Label category "How-To Guides"
- Users search "pricing" → Ensure "Pricing" in nav, not "Plans" or "Subscription"
3. **Descriptive breadcrumbs**:
- Bad: "Home > Section 1 > Page 3" → No meaning
- Good: "Home > Developer Docs > API Reference" → Clear path
4. **Preview text**: Show snippet of content under link
- Navigation item: "API Reference" + "Complete list of endpoints and parameters"
### Findability Metrics
**Key metrics to track**:
1. **Time to find**: How long to locate content?
- **Target**: <30 sec for simple tasks, <2 min for complex
- **Measurement**: Task completion time in usability tests
2. **Success rate**: % of users who find content?
- **Target**: ≥70% (tree test), ≥80% (live site with search)
- **Measurement**: Tree test results, task success in usability tests
3. **Search vs. browse**: Do users search or navigate?
- **Good**: 40-60% browse, 40-60% search (both work)
- **Bad**: 90% search (navigation broken), 90% browse (search broken)
- **Measurement**: Analytics (search usage %, nav click-through)
4. **Search refinement rate**: % of searches that are refined?
- **Target**: <30% (users find on first search)
- **Bad**: >50% (users search, refine, search again → poor results)
- **Measurement**: Analytics (queries per session)
5. **Bounce rate by entry point**: % leaving immediately?
- **Target**: <40% for landing pages
- **Bad**: >60% (users don't find what they expected)
- **Measurement**: Analytics (bounce rate by page)
6. **Navigation abandonment**: % who start navigating, then leave?
- **Target**: <20%
- **Bad**: >40% (users get lost, give up)
- **Measurement**: Analytics (drop-off in navigation funnels)
### Search vs. Navigation Trade-offs
**When search is preferred**:
- Large content sets (>5000 items)
- Users know exactly what they want ("lookup" mode)
- Diverse content types (hard to categorize consistently)
**When navigation is preferred**:
- Smaller content sets (<500 items)
- Users browsing, exploring ("discovery" mode)
- Hierarchical domains (clear parent-child relationships)
**Best practice**: Offer BOTH
- Navigation for discovery, context, exploration
- Search for lookup, speed, known-item finding
**Optimizing search**:
- **Autocomplete**: Suggest as user types
- **Filters**: Narrow results by category, date, type
- **Best bets**: Featured results for common queries
- **Zero-results page**: Suggest alternatives, show popular content
**Optimizing navigation**:
- **Clear labels**: Match user vocabulary (card sort insights)
- **Faceted filters**: Browse + filter combination
- **Related links**: Help users discover adjacent content
- **Breadcrumbs**: Show path, enable backtracking
---
## 5. Advanced Topics
### Mental Models & User Research
**Mental model**: User's internal representation of how system works
**Why it matters**: Navigation should match user's mental model, not company's org chart
**Researching mental models**:
1. **Card sorting**: Reveals how users group/label content
2. **User interviews**: Ask "How would you organize this?" "What would you call this?"
3. **Tree testing**: Validates if proposed structure matches mental model
4. **First-click testing**: Where do users expect to find X?
**Common mismatches**:
- **Company thinks**: "Features" (technical view)
- **Users think**: "What can I do?" (task view)
- **Solution**: Rename to task-based labels ("Create Report", "Share Dashboard")
**Example**: SaaS product
- **Internal (wrong)**: "Modules" → "Synergistic Solutions" → "Widget Management"
- **User mental model (right)**: "Features" → "Reporting" → "Custom Reports"
### Cross-Cultural IA
**Challenge**: Different cultures have different categorization preferences
**Examples**:
- **Alphabetical**: Works for Latin scripts, not ideographic (Chinese, Japanese)
- **Color coding**: Red = danger (Western), Red = luck (Chinese)
- **Icons**: Mailbox icon = email (US), doesn't translate (many countries have different mailbox designs)
**Strategies**:
1. **Localization testing**: Card sort with target culture users
2. **Avoid culturally-specific metaphors**: "Home run", "touchdown" (US sports)
3. **Simple, universal labels**: "Home", "Search", "Help" (widely understood)
4. **Icons + text**: Don't rely on icons alone
### IA Governance
**Problem**: Taxonomy degrades over time without maintenance
**Governance framework**:
1. **Roles**:
- **Content owner**: Publishes content, assigns categories/tags
- **Taxonomy owner**: Maintains category structure, adds/removes categories
- **IA steward**: Monitors usage, recommends improvements
2. **Processes**:
- **Quarterly review**: Check taxonomy usage, identify issues
- **Change request**: How to propose new categories or restructure
- **Deprecation**: Process for removing outdated categories
- **Tag moderation**: Review user-generated tags, merge synonyms
3. **Metrics to monitor**:
- % content in "Other" or "Uncategorized" (should be <5%)
- Empty categories (no content) — remove or consolidate
- Oversized categories (>50% of content) — split into subcategories
4. **Tools**:
- CMS with taxonomy management
- Analytics to track usage
- Automated alerts (e.g., "Category X has no content")
### Personalization & Dynamic IA
**Concept**: Navigation adapts to user
**Approaches**:
1. **Audience-based**: Show different nav for different user types
- "For Developers", "For Marketers", "For Executives"
2. **History-based**: Prioritize recently visited or frequently used
- "Recently Viewed", "Your Favorites"
3. **Context-based**: Show nav relevant to current task
- "Related Articles", "Next Steps"
4. **Adaptive search**: Results ranked by user's past behavior
**Caution**: Don't over-personalize
- Users need consistency to build mental model
- Personalization should augment, not replace, standard navigation
### IA for Voice & AI Interfaces
**Challenge**: Traditional visual hierarchy doesn't work for voice
**Strategies**:
1. **Flat structure**: No deep nesting (can't show menu)
2. **Natural language categories**: "Where can I find information about X?" vs. "Navigate to Category > Subcategory"
3. **Conversational**: "What would you like to do?" vs. "Select option 1, 2, or 3"
4. **Context-aware**: Remember user's previous question, continue conversation
**Example**:
- **Web**: Home > Products > Electronics > Phones
- **Voice**: "Show me phones" → "Here are our top phone options..."
---
## Summary
**Card sorting** reveals user mental models through similarity matrices, dendrograms, and consensus scores. Outliers indicate unclear content.
**Taxonomy design** follows MECE principle (mutually exclusive, collectively exhaustive). Use faceted classification for scale, controlled vocabulary for consistency, and plan for evolution.
**Navigation optimization** balances breadth (many choices) vs. depth (many clicks). Optimal: 5-9 items per level, 3-4 levels deep. Progressive disclosure reduces initial complexity.
**Information scent** guides users with clear labels, trigger words, and descriptive breadcrumbs. Track findability metrics: time to find (<30 sec), success rate (≥70%), search vs. browse balance (40-60% each).
**Advanced techniques** include mental model research (card sort, interviews), cross-cultural adaptation, governance frameworks, personalization, and voice interface design.
**The goal**: Users can predict where information lives and find it quickly, regardless of access method.

View File

@@ -0,0 +1,395 @@
# Information Architecture Templates
Quick-start templates for content audits, card sorting, sitemaps, tree testing, and taxonomy design.
---
## Template 1: Content Audit
**When to use**: Before redesigning navigation, need inventory of existing content
### Content Audit Spreadsheet Template
| URL/ID | Title | Content Type | Category | Metadata | Last Updated | Owner | Status | Action | Notes |
|--------|-------|--------------|----------|----------|--------------|-------|--------|--------|-------|
| /about | About Us | Page | Company | meta-desc, img | 2024-01-15 | Marketing | Keep | Update copy | Outdated team bios |
| /api/auth | Authentication API | Docs | Developer | tags: auth, jwt | 2024-06-01 | Eng | Keep | None | Current |
| /blog/old-post | Old Post | Article | Blog | tags: news | 2020-03-10 | Marketing | **Remove** | Archive | 4 years old, irrelevant |
| /products/widget-a | Widget A | Product | Products | price, sku | 2024-05-20 | Sales | Keep | Recategorize | Should be in "Tools" not "Products" |
| /help/faq-duplicate | FAQ | Help | Support | tags: help | 2023-12-01 | Support | **Merge** | With /help/faq | Duplicate content |
### Audit Analysis Template
**Total Content Items**: [Count]
**By Type**:
- Pages: [Count] ([%])
- Docs: [Count] ([%])
- Articles: [Count] ([%])
- Products: [Count] ([%])
- Other: [Count] ([%])
**By Status**:
- Keep as-is: [Count] ([%])
- Update needed: [Count] ([%])
- Recategorize: [Count] ([%])
- Merge: [Count] ([%])
- Remove/Archive: [Count] ([%])
**Quality Issues**:
- Outdated (>2 years): [Count] items
- Missing metadata: [Count] items
- Duplicates: [Count] items
- Broken links: [Count] items
- Inconsistent labeling: [Count] items
**Current Performance** (from analytics):
- Most visited pages: [List top 5]
- Least visited pages: [List bottom 5]
- High bounce rate pages (>60%): [List]
- Search queries with no results: [List top 10]
- Common navigation paths: [List top 3]
**Key Findings**:
1. [Finding 1: e.g., "40% of content is outdated"]
2. [Finding 2: e.g., "Users search for 'pricing' but it's buried 4 levels deep"]
3. [Finding 3: e.g., "15% of content is duplicated across sections"]
**Recommendations**:
1. [Recommendation 1]
2. [Recommendation 2]
3. [Recommendation 3]
---
## Template 2: Card Sorting Study
**When to use**: Understanding user mental models for content organization
### Card Sorting Plan Template
**Study Type**: [Open / Closed / Hybrid]
**Research Question**: "How do [target users] naturally group and label [content type]?"
**Participants**:
- Target: [Number, e.g., "20 users"]
- Recruitment criteria: [e.g., "Active customers who use product 2+ times/week"]
- Incentive: [e.g., "$25 gift card"]
**Cards** ([Number] total):
1. [Card 1 name/title]
2. [Card 2 name/title]
3. [Card 3 name/title]
...
[List all cards — aim for 30-60 cards, representative sample of content]
**Pre-Defined Categories** (for Closed/Hybrid only):
1. [Category 1]
2. [Category 2]
3. [Category 3]
...
**Tool**: [e.g., "OptimalSort", "UserZoom", "Miro"]
**Instructions for Participants**:
```
You'll see [X] cards representing different [content/features/topics] on our [website/app].
Your task:
1. Read each card
2. Group cards that belong together
3. Name each group you create
4. You can create as many groups as you need (typically 5-10 groups work well)
There are no right or wrong answers. We want to understand how YOU think about organizing this content.
Time estimate: 15-20 minutes
```
### Card Sort Analysis Template
**Participation**:
- Completed: [X] / [Y] participants ([Z]% completion rate)
- Average time: [X minutes]
**Category Analysis**:
| Category Label | # Users Who Created It | % Agreement | Cards Most Often Included | Alternative Labels Used |
|----------------|------------------------|-------------|---------------------------|------------------------|
| Getting Started | 18 / 20 | 90% | "Sign Up", "First Login", "Quick Start" | "Onboarding", "Setup" |
| Reports | 16 / 20 | 80% | "Analytics", "Dashboard", "Metrics" | "Insights", "Data" |
| Settings | 19 / 20 | 95% | "Profile", "Preferences", "Account" | "Configuration", "Options" |
**Dendogram Insights** (hierarchical clustering):
- **Strong clusters** (>70% agreement): [List categories with high consensus]
- **Weak clusters** (<50% agreement): [List categories with low consensus — these need clarification]
- **Outlier cards**: [Cards that don't fit anywhere — may need reconsideration]
**Labeling Insights**:
- **Most common labels**: [e.g., "Help" (15 users), "Support" (12 users), "Assistance" (3 users)]
- **Terminology conflicts**: [e.g., "Users say 'Reports' but we call them 'Analytics'"]
- **Jargon to avoid**: [e.g., "Synergistic Solutions" — 0 users used this term]
**Key Findings**:
1. [Finding 1: e.g., "Users group by task, not by product feature"]
2. [Finding 2: e.g., "'Settings' is universally understood (95% agreement)"]
3. [Finding 3: e.g., "'Modules' label confuses users — split between 'Features' and 'Tools'"]
**Recommended Categories** (based on card sort):
1. [Category 1] — [Brief description]
2. [Category 2] — [Brief description]
3. [Category 3] — [Brief description]
...
---
## Template 3: Sitemap & Navigation Structure
**When to use**: Documenting information hierarchy after card sort analysis
### Sitemap Template
```
Homepage
├── [Category 1]
│ ├── [Subcategory 1.1]
│ │ ├── [Page 1.1.1]
│ │ └── [Page 1.1.2]
│ ├── [Subcategory 1.2]
│ │ ├── [Page 1.2.1]
│ │ └── [Page 1.2.2]
│ └── [Subcategory 1.3]
├── [Category 2]
│ ├── [Subcategory 2.1]
│ ├── [Subcategory 2.2]
│ └── [Subcategory 2.3]
├── [Category 3]
│ ├── [Page 3.1]
│ ├── [Page 3.2]
│ └── [Page 3.3]
└── [Category 4]
├── [Subcategory 4.1]
│ ├── [Page 4.1.1]
│ └── [Page 4.1.2]
└── [Subcategory 4.2]
```
### Example: E-commerce Sitemap
```
Homepage
├── Shop by Category (Electronics, Clothing, Home & Garden)
├── Shop by Occasion (Daily Essentials, Special Occasions, Gifts)
├── Deals & Promotions (Today's Deals, Clearance, Bundles)
└── Support (Help Center, Track Order, Returns, Contact Us)
```
### Navigation Specification Template
**Primary Navigation** (Top-level categories):
- [Category 1]: [Description, user goal, target audience]
- [Category 2]: [Description, user goal, target audience]
- [Category 3]: [Description, user goal, target audience]
**Secondary Navigation** (Utility links):
- [Link 1]: [e.g., "Search", "Cart", "Account"]
- [Link 2]: [e.g., "Help", "Contact"]
**Faceted Filters** (for browse/search):
- **Facet 1**: [Name] — Options: [Option A, Option B, Option C]
- **Facet 2**: [Name] — Options: [Option A, Option B, Option C]
- **Facet 3**: [Name] — Options: [Option A, Option B, Option C]
**Breadcrumbs**: [Show path from homepage to current page]
- Example: `Home > [Category] > [Subcategory] > [Current Page]`
**Related Links**: [Contextual links on each page]
- "Related Articles"
- "You Might Also Like"
- "Next Steps"
**Search**:
- Placement: [e.g., "Global header, every page"]
- Features: [e.g., "Autocomplete, filters, recent searches"]
- Best bets: [Featured results for common queries]
---
## Template 4: Tree Testing Script
**When to use**: Validating navigation structure before building
### Tree Test Plan Template
**Research Question**: "Can users find [X] in our proposed navigation structure?"
**Participants**: [20-50 users matching target audience]
**Navigation Tree** (text-only structure):
```
Home
Getting Started
Sign Up
First Login
Quick Start Guide
Features
Reports
Dashboards
Alerts
Settings
Profile
Preferences
Billing
Support
Help Center
Contact Us
FAQ
```
**Tasks** (8-12 tasks typical):
1. **Task 1**: "You want to [user goal]. Where would you find [content]?"
- **Correct destination**: [Path, e.g., "Features > Reports"]
- **Why this task**: [e.g., "Most common user action"]
2. **Task 2**: "You need to [user goal]. Where would you look?"
- **Correct destination**: [Path]
- **Why this task**: [Rationale]
3. **Task 3**: [Task description]
- **Correct destination**: [Path]
- **Why this task**: [Rationale]
[Continue for 8-12 tasks covering key content/features]
### Tree Test Results Template
**Overall Performance**:
- **Success rate**: [X]% (Target: ≥70%)
- **Directness**: [X]× optimal path (Target: ≤1.5×)
- **Average time**: [X] seconds
**Task-Level Results**:
| Task | Success Rate | Directness | Avg Time | Top Incorrect Paths | Notes |
|------|--------------|------------|----------|---------------------|-------|
| Task 1 | 85% | 1.2× | 12 sec | None | Good |
| Task 2 | 45% | 2.5× | 35 sec | "Settings > Profile" (30%), "Support > Help" (25%) | **PROBLEM**: Users confused by label |
| Task 3 | 75% | 1.4× | 18 sec | "Features > Dashboards" (20%) | Minor issue |
**Problem Areas**:
| Issue | Evidence | Impact | Recommendation |
|-------|----------|--------|----------------|
| "Reports" label unclear | Task 2 only 45% success, users looked in "Dashboards" and "Settings" | High | Rename to "Analytics" or merge with "Dashboards" |
| "Settings" too broad | Tasks 4 & 5 both low success (50%, 55%) | Medium | Split into "Account Settings" and "App Preferences" |
| Deep nesting in "Support" | Task 7: 65% success, 2.8× directness | Medium | Flatten: move "FAQ" to top level |
**Recommended Changes**:
1. [Change 1: e.g., "Rename 'Reports' to 'Analytics'"]
2. [Change 2: e.g., "Split 'Settings' into 'Account' and 'Preferences'"]
3. [Change 3: e.g., "Move 'FAQ' from Support > Help Center > FAQ to top-level Support > FAQ"]
**Next Steps**:
- [ ] Implement recommended changes
- [ ] Re-run tree test on updated structure
- [ ] Proceed to wireframes once success rate ≥70% on all tasks
---
## Template 5: Taxonomy & Metadata Schema
**When to use**: Designing classification systems, tags, and content attributes
### Taxonomy Design Template
**Content Type**: [e.g., "Product", "Article", "Documentation Page"]
**Primary Taxonomy** (Hierarchical categories):
```
[Top-Level Category 1]
├── [Sub-category 1.1]
├── [Sub-category 1.2]
└── [Sub-category 1.3]
[Top-Level Category 2]
├── [Sub-category 2.1]
└── [Sub-category 2.2]
[Top-Level Category 3]
```
**Facets** (Orthogonal dimensions for filtering):
- **Facet 1**: [Name] — Values: [Value A, Value B, Value C]
- **Facet 2**: [Name] — Values: [Value A, Value B, Value C]
- **Facet 3**: [Name] — Values: [Value A, Value B, Value C]
**Tags** (Flexible, user-defined or preset):
- Preset tags: [List of controlled vocabulary tags]
- User-generated tags: [Yes/No]
- Tag moderation: [Yes/No, process]
**Metadata Schema**:
| Field | Type | Required | Values/Format | Purpose |
|-------|------|----------|---------------|---------|
| Title | Text | Yes | Plain text, max 60 chars | Display in nav, SEO |
| Description | Text | No | Plain text, max 160 chars | SEO meta description |
| Category | Select | Yes | [List of categories] | Primary classification |
| Tags | Multi-select | No | [Controlled vocabulary] | Cross-category findability |
| Author | Text | Yes | Name or ID | Attribution, filtering |
| Last Updated | Date | Yes | YYYY-MM-DD | Freshness indicator |
| Audience | Select | Yes | [e.g., "Beginner", "Advanced"] | Personalization |
| Content Type | Select | Yes | [e.g., "Tutorial", "Reference", "FAQ"] | Filtering, display |
| Status | Select | Yes | [e.g., "Draft", "Published", "Archived"] | Governance |
### Example: Documentation Taxonomy
**Primary Taxonomy**:
```
Getting Started
├── Installation
├── Configuration
└── First App
Guides
├── How-To Guides
├── Tutorials
└── Best Practices
Reference
├── API Documentation
├── CLI Reference
└── Configuration Files
Troubleshooting
├── Common Issues
├── Error Messages
└── FAQ
```
**Facets**:
- **Product Version**: v1.0, v2.0, v3.0
- **Difficulty Level**: Beginner, Intermediate, Advanced
- **Time to Complete**: <10 min, 10-30 min, >30 min
**Metadata** (for each doc page):
- Title, Description, Category, Tags, Author, Last Updated, Audience, Content Type, Product Version, Difficulty, Est. Time
---
## Quick Reference: When to Use Each Template
| Template | Use Case | Deliverable |
|----------|----------|-------------|
| **Content Audit** | Before redesign, need inventory | Spreadsheet with all content, status, recommendations |
| **Card Sorting** | Don't know how users categorize content | Category labels, groupings, terminology insights |
| **Sitemap** | After card sort, document structure | Visual hierarchy diagram |
| **Tree Testing** | Validate navigation before building | Success rates, problem areas, iteration recommendations |
| **Taxonomy & Metadata** | Design classification system | Taxonomy structure, metadata schema, controlled vocabulary |