Use when validating product assumptions before building, discovering unmet user needs, understanding customer problems and workflows, testing concepts or positioning, researching target markets, identifying jobs-to-be-done and hiring triggers, uncovering pain points and workarounds, or when users mention user research, customer interviews, surveys, discovery interviews, validation studies, or voice of customer.
12k tokens
context cost
the whole folder, loaded on every use
4
files
instructions only
0
copies elsewhere
how many repositories repackaged it
275
stars on the repo
on the repository, not the skill itself
Install
one command, takes just this skill from the repository
Discovery Interviews & Surveys provide structured approaches to learn from users while avoiding common biases (leading questions, confirmation bias, selection bias).
Key components:
Interview guides: Open-ended questions that reveal problems and context
Survey instruments: Scaled questions for quantitative validation at scale
JTBD probes: Questions focused on hiring/firing triggers and desired outcomes
Bias-avoidance techniques: Past behavior focus, "show me" requests, avoiding hypotheticals
"Would you pay $49/month for a tool that automatically backs up your files?"
Good interview approach (behavior-focused, problem-discovery):
"Tell me about the last time you lost important files. What happened?"
"What have you tried to prevent data loss? How's that working?"
"Walk me through your current backup process. Show me if possible."
"What would need to change for you to invest time/money in better backup?"
Result: Learn about actual problems, current solutions, willingness to change—not hypothetical preferences.
Workflow
Copy this checklist and track your progress:
Discovery Research Progress:
- [ ] Step 1: Define research objectives and hypotheses
- [ ] Step 2: Identify target participants
- [ ] Step 3: Choose research method (interviews, surveys, or both)
- [ ] Step 4: Design research instruments
- [ ] Step 5: Conduct research and collect data
- [ ] Step 6: Analyze findings and extract insights
Step 1: Define research objectives
Specify what you're trying to learn, key hypotheses to test, success criteria for research, and decision to be informed. See Common Patterns for typical objectives.
Step 2: Identify target participants
Define participant criteria (demographics, behaviors, firmographics), sample size needed, recruitment strategy, and screening questions. For sampling strategies, see resources/methodology.md.
Step 3: Choose research method
Based on objective and constraints:
For deep problem discovery (5-15 participants) → Use resources/template.md for in-depth interviews
For concept testing at scale (50-200+ participants) → Use resources/template.md for quantitative validation
For JTBD research → Use resources/methodology.md for switch interviews
For mixed methods → Interviews for discovery, surveys for validation
Step 4: Design research instruments
Create interview guide or survey with bias-avoidance techniques. Use resources/template.md for structure. Avoid leading questions, focus on past behavior, use "show me" requests. For advanced question design, see resources/methodology.md.
Step 5: Conduct research
Execute interviews (record with permission, take notes) or distribute surveys (pilot test first). Use proper techniques (active listening, follow-up probes, silence for thinking). See Guardrails for critical requirements.
Step 6: Analyze findings
For interviews: thematic coding, affinity mapping, quote extraction. For surveys: statistical analysis, cross-tabs, open-end coding. Create insights document with evidence. Self-assess using resources/evaluators/rubric_discovery_interviews_surveys.json. Minimum standard: Average score ≥ 3.5.
Common Patterns
Pattern 1: Problem Discovery Interviews
Objective: Understand user pain points and current workflows
Approach: 8-12 in-depth interviews, open-ended questions, focus on past behavior and actual solutions
Key questions: "Tell me about the last time...", "Walk me through...", "What have you tried?", "How's that working?"
Output: Problem themes, frequency estimates, current workarounds, willingness to change
Example: B2B SaaS discovery—interview potential customers about current tools and pain points
Pattern 2: Jobs-to-be-Done Research
Objective: Identify why users "hire" products and what triggers switching
Approach: Switch interviews with recent adopters or switchers, focus on timeline and context
Key questions: "What prompted you to look?", "What alternatives did you consider?", "What almost stopped you?", "What's different now?"
Key questions: Varies by current focus (new features, onboarding, expansion, retention)
Output: Continuous insight feed, early problem detection, relationship building
Example: Product team does 3-5 customer calls weekly, logs insights in shared doc
Guardrails
Critical requirements:
Avoid leading questions: Don't telegraph the "right" answer. Bad: "Don't you think our UI is confusing?" Good: "Walk me through using this feature. What happened?"
Focus on past behavior, not hypotheticals: What people did reveals truth; what they say they'd do is often wrong. Bad: "Would you use this feature?" Good: "Tell me about the last time you needed to do X."
Use "show me" not "tell me": Actual behavior > described behavior. Ask to screen-share, demonstrate current workflow, show artifacts (spreadsheets, tools).
Recruit right participants: Screen carefully. Wrong participants = wasted time. Define inclusion/exclusion criteria, use screening survey.
Sample size appropriate for method: Interviews: 5-15 for themes to emerge. Surveys: 100+ for statistical significance, 30+ per segment if comparing.
Avoid confirmation bias: Actively look for disconfirming evidence. If 9/10 interviews support hypothesis, focus heavily on the 1 that doesn't.
Record and transcribe (with permission): Memory is unreliable. Record interviews, transcribe for analysis. Take notes as backup.
Analyze systematically: Don't cherry-pick quotes that support preferred conclusion. Use thematic coding, count themes, present contradictory evidence.
Common pitfalls:
❌ Asking "would you" questions: Hypotheticals are unreliable. Focus on "have you", "tell me about when", "show me"
❌ Small sample statistical claims: "80% of users want feature X" from 5 interviews is not valid. Interviews = themes, surveys = statistics
❌ Selection bias: Interviewing only enthusiasts or only detractors skews results. Recruit diverse sample
❌ Ignoring non-verbal cues: Hesitation, confusion, workarounds during "show me" reveal truth beyond words
❌ Stopping at surface answers: First answer is often rationalization. Follow up: "Tell me more", "Why did that matter?", "What else?"