19 KiB
UX Research
User research methodologies and insights synthesis
Research Methods Skill
Comprehensive guide to UX research methodologies and when to use each method
This skill codifies best practices from hundreds of successful UX research projects across discovery, validation, and optimization phases.
Core Principles
- Match Method to Question: Choose research methods based on what you need to learn, not what's trendy
- Triangulate: Use multiple methods to validate findings and reduce bias
- Include Users Early and Often: Research throughout the product lifecycle, not just at the beginning
- Balance Qual and Quant: Qualitative explores "why," quantitative measures "how much"
- Ethical Research Always: Respect participant time, privacy, and dignity
Research Method Selection Framework
The 2x2 Framework: Attitudinal vs. Behavioral, Qualitative vs. Quantitative
QUALITATIVE QUANTITATIVE
(Why/How) (How many/much)
┌─────────────────────┬──────────────────────────┐
│ │ │
ATTITUDINAL │ • Interviews │ • Surveys │
(What people │ • Focus Groups │ • Card Sorting (quant) │
say) │ • Diary Studies │ • A/B Tests (attitudes) │
│ │ • Analytics (stated) │
├─────────────────────┼──────────────────────────┤
│ │ │
BEHAVIORAL │ • Usability Tests │ • Analytics │
(What people │ • Field Studies │ • A/B Tests │
do) │ • Contextual Inq. │ • Heatmaps/Click tracks │
│ • Task Analysis │ • Performance Metrics │
│ │ │
└─────────────────────┴──────────────────────────┘
When to Use Each Quadrant
Attitudinal + Qualitative: Understanding motivations, perceptions, mental models
- "Why do users prefer X over Y?"
- "What do users think about this concept?"
Attitudinal + Quantitative: Measuring opinions at scale
- "How satisfied are users with feature X?"
- "How many users would pay for premium tier?"
Behavioral + Qualitative: Understanding how users interact and why
- "How do users currently accomplish task X?"
- "What workarounds do users create?"
Behavioral + Quantitative: Measuring what users do at scale
- "What percentage of users complete onboarding?"
- "Which features are used most frequently?"
Research Methods by Purpose
1. GENERATIVE RESEARCH (Discovery & Exploration)
Goal: Understand user needs, contexts, and problems before defining solutions
User Interviews
Best for: Deep understanding of motivations, attitudes, experiences
When to use:
- Beginning of project to understand problem space
- Exploring new market or user segment
- Understanding "why" behind behaviors
Pros:
- Rich, deep insights
- Flexibility to explore unexpected topics
- Can uncover unarticulated needs
Cons:
- Time-consuming (typically 45-90 min per interview)
- Small sample size (5-12 participants)
- Susceptible to social desirability bias
Sample size: 5-12 participants per user segment
Typical questions:
- "Tell me about the last time you..."
- "Walk me through how you..."
- "What's most important to you when..."
Contextual Inquiry / Field Studies
Best for: Observing users in their natural environment
When to use:
- Understanding workflow in complex environments
- Observing actual behavior (not self-reported)
- Discovering workarounds and adaptations
Pros:
- Observe real behavior in context
- See environmental factors affecting use
- Discover unarticulated needs
Cons:
- Very time-intensive
- Logistically complex (travel, access)
- Can be intrusive to participants
Sample size: 4-8 site visits
Process:
- Shadow user in their environment
- Observe tasks and workflows
- Ask questions about what you observe
- Document context (photos, notes)
Diary Studies
Best for: Understanding behaviors and experiences over time
When to use:
- Behaviors that occur sporadically
- Understanding context of use over time
- Capturing in-the-moment experiences
Pros:
- Captures real-time experiences (less recall bias)
- Shows patterns over time
- Less researcher time than shadowing
Cons:
- Participant burden (compliance issues)
- Self-reported data
- Can take weeks to complete
Sample size: 10-20 participants over 1-4 weeks
Tools: Mobile apps (dscout, Indeemo), photo journals, SMS surveys
Surveys (Exploratory)
Best for: Identifying patterns and priorities across large populations
When to use:
- Need to understand frequency of behaviors
- Want to segment users
- Validate interview findings at scale
Pros:
- Large sample sizes
- Statistically significant results
- Cost-effective per response
Cons:
- Can't explore "why" deeply
- Response bias
- No opportunity to clarify answers
Sample size: Minimum 100 for basic analysis, 300+ for segmentation
Question types:
- Multiple choice for quantification
- Likert scales for attitudes
- Open-ended for unexpected insights (limited)
2. EVALUATIVE RESEARCH (Testing & Validation)
Goal: Test designs, concepts, or products to identify issues and validate solutions
Usability Testing (Moderated)
Best for: Identifying usability issues and understanding user behavior with product
When to use:
- Testing prototypes or live products
- Evaluating specific flows or features
- Before major product launches
Pros:
- Identifies specific usability issues
- Can probe to understand "why"
- Flexible to explore unexpected findings
Cons:
- Artificial environment (not real context)
- Moderator bias possible
- Time-consuming
Sample size: 5 users per user segment (finds ~85% of issues per Nielsen)
Process:
- Give user realistic task scenarios
- Ask them to think aloud
- Observe struggles and successes
- Measure time, errors, completion
Metrics:
- Task completion rate
- Time on task
- Error rate
- Satisfaction ratings (post-task)
Usability Testing (Unmoderated Remote)
Best for: Quick feedback on specific tasks at scale
When to use:
- Need results quickly
- Testing simple, linear flows
- Want larger sample size
Pros:
- Fast turnaround (hours, not weeks)
- Larger sample sizes
- Lower cost per participant
Cons:
- Can't ask follow-up questions
- Limited to simple tasks
- No way to clarify misunderstandings
Sample size: 15-30 participants
Tools: UserTesting, Maze, UserZoom
A/B Testing
Best for: Comparing two design alternatives with real users
When to use:
- Have two viable design options
- Need to measure impact on key metrics
- Have sufficient traffic for significance
Pros:
- Measures real behavior (not stated preferences)
- Statistically significant results
- Removes opinion-based decisions
Cons:
- Requires traffic volume
- Can only test one variable at a time
- Doesn't explain "why" variant won
Sample size: Depends on baseline conversion and desired lift (use calculator)
Metrics: Conversion rate, click-through rate, time on page, revenue
Concept Testing
Best for: Validating ideas before building them
When to use:
- Early stage, before detailed design
- Choosing between multiple concepts
- Validating value proposition
Pros:
- Fail fast and cheap
- Can test multiple concepts
- Provides direction for design
Cons:
- Reactions to concept ≠ actual usage
- Risk of overvaluing stated intent
- Hard to imagine new paradigms
Sample size: 8-15 for qualitative, 100+ for quantitative
Methods: Show concept (sketches, storyboards, mockups), gather reactions
Card Sorting
Best for: Testing or creating information architecture
When to use:
- Designing navigation structure
- Organizing content or features
- Understanding mental models of categorization
Types:
- Open: Users create and name categories
- Closed: Users sort into predefined categories
- Hybrid: Mix of both
Pros:
- User-centered IA
- Quantifiable results
- Easy to conduct remotely
Cons:
- Decontextualized from real use
- Only tests grouping, not findability
- Assumes users know all items
Sample size: 15-30 for open, 30+ for closed
Tools: Optimal Workshop, Maze, UserZoom
Tree Testing
Best for: Validating navigation structure
When to use:
- After creating IA (validates card sorting)
- Testing if users can find content
- Before visual design
Pros:
- Tests findability directly
- Quantifiable results
- Fast and cost-effective
Cons:
- No visual design context
- Text-only (misses visual cues)
- Limited to tree structures
Sample size: 50-100 participants
Metrics: Success rate, directness, time
3. ANALYTICS & BEHAVIORAL DATA
Web/Product Analytics
Best for: Understanding actual user behavior at scale
When to use:
- Identifying where users drop off
- Measuring feature adoption
- Tracking key metrics over time
- Generating hypotheses for research
Pros:
- Real behavior, not self-reported
- Large sample sizes
- Continuous monitoring
Cons:
- Doesn't explain "why"
- Privacy concerns
- Can be misinterpreted without context
Key Metrics:
- Acquisition: Where users come from
- Activation: Do they complete onboarding?
- Engagement: How often do they return?
- Retention: Do they stick around?
- Revenue: Do they pay?
Tools: Google Analytics, Mixpanel, Amplitude, Heap
Heatmaps & Session Recordings
Best for: Visualizing where users click, scroll, and focus attention
When to use:
- Optimizing page layouts
- Understanding confusion points
- Identifying ignored content
Pros:
- Visual and intuitive
- Shows aggregate patterns
- Can identify rage clicks (frustration)
Cons:
- Descriptive, not explanatory
- Can be over-interpreted
- Privacy concerns
Tools: Hotjar, Crazy Egg, FullStory
4. SPECIALIZED METHODS
Accessibility Testing
Best for: Ensuring product works for users with disabilities
When to use:
- Required for compliance (WCAG, ADA)
- Serving diverse user base
- Ethical product development
Methods:
- Automated testing (axe, WAVE, Lighthouse)
- Manual testing with assistive tech
- User testing with people with disabilities
Sample size: 3-5 users per disability type
Competitive Analysis
Best for: Understanding market landscape and best practices
When to use:
- Entering new market
- Benchmarking against competitors
- Identifying feature gaps
Process:
- Identify competitors
- Analyze key features, flows, pricing
- Conduct comparative usability testing
- Document strengths and weaknesses
Heuristic Evaluation
Best for: Expert review against usability principles
When to use:
- Quick assessment without users
- Budget/time constraints
- Before user testing to fix obvious issues
Heuristics (Nielsen's 10):
- Visibility of system status
- Match between system and real world
- User control and freedom
- Consistency and standards
- Error prevention
- Recognition rather than recall
- Flexibility and efficiency of use
- Aesthetic and minimalist design
- Help users recognize, diagnose, and recover from errors
- Help and documentation
Evaluators: 3-5 UX experts
Research Method Selection Guide
By Project Stage
Discovery / Problem Definition:
- User interviews
- Field studies / Contextual inquiry
- Diary studies
- Surveys (exploratory)
- Competitive analysis
Ideation / Concept Validation:
- Concept testing
- Participatory design workshops
- Card sorting (for IA)
- Surveys (concept preference)
Design / Prototype Testing:
- Usability testing (moderated)
- Tree testing
- First-click testing
- Heuristic evaluation
Pre-Launch Validation:
- Usability testing (unmoderated)
- Accessibility testing
- Beta testing
- A/B testing (if redesign)
Post-Launch Optimization:
- Analytics
- A/B testing
- Surveys (satisfaction)
- Session recordings
- Ongoing usability testing
By Research Question Type
"What do users need?" → Interviews, field studies, diary studies
"Will users understand this concept?" → Concept testing, first-click tests
"Can users complete this task?" → Usability testing, task analysis
"How many users experience this?" → Surveys, analytics
"Which design performs better?" → A/B testing, preference testing
"How satisfied are users?" → Surveys (NPS, CSAT, SUS), interviews
"How do we organize content?" → Card sorting, tree testing
"Is this accessible?" → Accessibility testing, assistive tech testing
Sample Size Guidelines
Qualitative Research
Interviews / Usability Tests: 5-8 per user segment
- Diminishing returns after 5 (per Nielsen)
- If multiple distinct segments, 5 per segment
- For very simple products: 3-5 total
- For complex products: 8-12 total
Field Studies: 4-8 site visits
- Very time-intensive
- Smaller sample acceptable
Diary Studies: 10-20 participants
- Higher dropout rate
- Need larger initial sample
Quantitative Research
Surveys:
- Exploratory: 100+ for basic insights
- Segmentation analysis: 300+
- Statistical modeling: 1000+
A/B Tests:
- Depends on baseline conversion rate and desired lift
- Typically need thousands of users per variant
- Use online calculator for specific estimates
Card Sorting:
- Open: 15-30
- Closed: 30-50
Tree Testing: 50-100
Mixed Methods Approach
Best practice: Combine qualitative and quantitative for robust insights
Example: Redesigning Checkout Flow
-
Discover (Qual): Interview 8 users about current checkout experience
- Identify pain points and needs
-
Validate (Quant): Survey 300 users on identified pain points
- Quantify how many experience each issue
-
Test (Qual): Usability test new design with 6 users
- Identify usability issues
-
Measure (Quant): A/B test new design with live traffic
- Measure impact on conversion
-
Understand (Qual): Interview users who abandoned new checkout
- Understand why some still fail
This creates a complete picture: qualitative explains, quantitative measures.
Common Research Mistakes to Avoid
1. Confirmation Bias
❌ Mistake: Only asking questions that support your hypothesis ✅ Fix: Include questions that could disprove your assumptions
2. Leading Questions
❌ Mistake: "Don't you think this design is confusing?" ✅ Fix: "What are your thoughts on this design?"
3. Hypothetical Questions
❌ Mistake: "Would you use this feature?" ✅ Fix: "Tell me about the last time you needed to [accomplish task]"
4. Treating Insights as Data
❌ Mistake: "5 users said X, so everyone must think X" ✅ Fix: "5/5 users experienced X (small sample, directional insight)"
5. Testing Too Late
❌ Mistake: Only testing after full build ✅ Fix: Test early with low-fidelity prototypes
6. Ignoring Negative Findings
❌ Mistake: Cherry-picking data that supports your design ✅ Fix: Report all findings, especially contradictory ones
7. Over-Researching
❌ Mistake: Endless research without action ✅ Fix: Define decision criteria upfront, research only to inform decision
8. Research Without Clear Objectives
❌ Mistake: "Let's do some user interviews and see what we find" ✅ Fix: "We need to understand [specific question] to decide [specific decision]"
Ethical Research Guidelines
Informed Consent
- Explain purpose of research clearly
- Explain how data will be used
- Explain recording and privacy practices
- Allow participant to opt out at any time
- Obtain explicit consent before proceeding
Privacy and Confidentiality
- Anonymize participant data
- Secure storage of recordings and notes
- Delete personal information after analysis
- Don't share recordings outside research team
- Follow GDPR/privacy regulations
Respectful Treatment
- Compensate participants fairly for time
- Don't make participants feel judged
- Accommodate accessibility needs
- Allow breaks and adjust pace
- Thank participants sincerely
Vulnerable Populations
- Extra care with children, elderly, disabled
- May require guardian consent
- Trauma-informed approach for sensitive topics
- Have resources ready if distress occurs
Research Planning Checklist
Before Research:
- Clear research questions defined
- Method(s) selected and justified
- Participant criteria specified
- Recruitment plan in place
- Materials prepared (guides, prototypes)
- Consent forms ready
- Recording setup tested
- Pilot test conducted
During Research:
- Follow protocol consistently
- Take detailed notes
- Record sessions (with consent)
- Note unexpected findings
- Debrief after each session
- Adjust if critical issues found
After Research:
- Analyze data systematically
- Identify patterns and themes
- Create actionable insights
- Prioritize recommendations
- Share findings with stakeholders
- Track implementation of recommendations
Quick Reference: Method Selection Matrix
| If you need to... | Use this method |
|---|---|
| Understand user needs before designing | Interviews, Field Studies |
| Test if users can complete a task | Usability Testing |
| Choose between two designs | A/B Testing, Preference Testing |
| Organize content/features | Card Sorting |
| Validate navigation structure | Tree Testing |
| Measure satisfaction | Surveys (SUS, NPS, CSAT) |
| Understand "why" behind analytics | Interviews, Session Recordings |
| Validate a new concept | Concept Testing |
| Track behavior over time | Diary Studies, Analytics |
| Quick feedback on prototype | Unmoderated Usability Testing |
| Understand expert perspective | Heuristic Evaluation |
| Ensure accessibility | Accessibility Testing |
Version: 1.0 Last Updated: January 2025 Success Rate: Based on industry best practices and research literature