User research

User Research Patterns

Research Method Selection

Method Matrix

Method	Attitudinal vs Behavioral	Qualitative vs Quantitative	When to Use
User interviews	Attitudinal	Qualitative	Discovery, deep motivations
Usability testing	Behavioral	Qualitative	Evaluate design, find friction
Surveys	Attitudinal	Quantitative	Validate, measure, segment
Card sorting	Attitudinal	Both	Information architecture
Diary studies	Behavioral	Qualitative	Longitudinal, in-context
A/B testing	Behavioral	Quantitative	Optimize specific decisions
Eye tracking	Behavioral	Quantitative	Visual attention patterns
Analytics review	Behavioral	Quantitative	Understand current behavior

Method Selection by Phase

Discover → Interviews, diary studies, ethnography
Define → Affinity mapping, synthesis, personas
Design → Card sorting, prototype testing
Validate → Usability testing, A/B tests
Post-launch → Surveys, analytics, retention analysis

User Interviews

Recruiting

• Sample size: 5 users per distinct segment (Nielsen's rule for usability)
• Recruiting channels: Existing users, user panel (Respondent.io, UserTesting), LinkedIn, Reddit
• Screening criteria: Write a screener with 5-8 questions to qualify participants

Interview Guide Template

Study: [Name]  Duration: 60 min  Moderator: ___

INTRO (5 min)
"Thanks for joining. I'm [name] from [company]. We're doing research to 
understand how people [problem area]. There are no right or wrong answers—
we want to understand your actual experience. This session will be recorded. 
Any questions before we start?"

WARM-UP (5 min)
- "Tell me about your role"
- "What does a typical day look like?"

CURRENT BEHAVIOR (20 min)
- "Tell me about the last time you [task related to problem]"
- "Walk me through exactly what you did, step by step"
- "What tools did you use?"
- "What was hard about that?"
- "What did you do when [specific friction point]?"

PROBE DEEPER (15 min)
- "Why did you do it that way?"
- "What would ideal look like?"
- "How often does this happen?"
- "Who else is involved?"

CONCEPT TESTING (10 min) [optional]
- "Here's something we're considering. What's your first reaction?"
- "What would you expect to happen when you [action]?"
- "Who would use this?"

WRAP-UP (5 min)
- "Is there anything else you think I should know?"
- "Is there anyone else you think I should talk to?"

Interview Best Practices

• Silence is ok — let it breathe, don't fill it
• Ask about past behavior, not hypothetical ("would you")
• Avoid leading: NOT "Would you say the checkout is confusing?" → "Walk me through what happened in checkout"
• 5 Whys: keep asking "why" to get to root motivation
• One question at a time — don't bundle multiple questions

Usability Testing

Moderated Usability Test Plan

Objective: Evaluate [specific design / flow] for [user goal]

Tasks:
1. "You just signed up. Show me how you'd [first task]."
   Success criteria: [what counts as success]
   Metrics: Time on task, errors, completion rate

2. "Now, [second task]."
   Success criteria: ___

Observation notes:
- Verbal think-aloud
- Body language / hesitation
- Errors and recovery attempts
- Exact quotes

Metrics to capture per task:
[ ] Completion rate (%)
[ ] Time on task (seconds)
[ ] Error count
[ ] Satisfaction (1-7 scale after each task)

SUS (System Usability Scale)

10-item survey, 5-point Likert, scored 0-100:

• 80.3: Excellent (A)

• 68-80.3: Good (B/C)
• 51-68: OK (D)
• < 51: Poor (F)

Questions alternate positive/negative phrasing to detect satisficing.

Card Sorting

Open Card Sort (Discovery)

Participants group cards and name categories themselves.

• Use: Information architecture discovery, navigation labels
• Tool: Optimal Workshop, Maze, physical index cards
• Cards: 30-40 items max, 15-20 participants

Closed Card Sort (Validation)

Participants sort cards into predefined categories.

• Use: Validate proposed structure, find misfits
• Cards: Same items, categories you've designed

Analysis

Dendogram: Shows clusters of items frequently sorted together
Similarity matrix: % of participants who grouped items together
Most common category labels → use as navigation terms

Affinity Mapping

Process

1. Collect raw observations — quotes, behaviors, pain points on sticky notes (one per note)
2. Silent sort — team members group similar notes without talking
3. Label clusters — write a label that captures the theme
4. Find meta-themes — group clusters into higher-level themes
5. Prioritize — vote on most important or most frequent themes

Note Format

[What happened]: "The user spent 3 minutes looking for the settings icon"
[Quote]: "I had no idea where to find my account info"
[Inference]: Users expect account controls in top-right corner

Diary Studies

Diary Study Protocol

• Duration: 1-4 weeks of participant self-logging
• Frequency: Daily or event-triggered entries
• Prompts:
  • "Describe what you just did related to [topic]"
  • "Take a screenshot and explain what you're seeing"
  • "Rate your frustration right now (1-10) and explain"
• Tools: WhatsApp group, Google Form, dscout, Notion form

Personas

Evidence-Based Persona Template

Name: [Fictional but realistic]
Photo: [Stock photo]
Role: [Job title]
Age: [Range]
Location: [Where they work/live]

Quote: "[Verbatim quote that captures their worldview]"

Goals:
• [Primary goal]
• [Secondary goal]

Frustrations:
• [Pain point 1 — backed by research]
• [Pain point 2]

Behaviors:
• Uses [tools] for [tasks]
• Checks [channels] for [information]

Decision process:
• Who influences their decisions?
• What criteria matter most?

How we help:
• [How your product maps to their goal]

Anti-pattern: Demographic Persona

Avoid personas based only on age/gender/income without behavioral grounding. A 45-year-old CFO and a 45-year-old nurse have very different jobs to be done.

Research Synthesis

Insight Statement Format

[User group] needs a way to [achieve goal] 
because [underlying motivation or barrier].

Example:
"Senior engineers need a way to communicate architecture decisions
to junior teammates because context switching to explain slows
down their deep work and creates repeated questions."

Opportunity → HMW Framing

Convert insights to "How Might We" questions:

• Insight: Users abandon checkout when asked for account creation
• HMW: "How might we let users complete purchase without creating an account?"