opencode-metis 0.1.0

package/opencode/skill/user-insight-synthesis/SKILL.md

@@ -0,0 +1,576 @@
+---
+name: user-insight-synthesis
+description: "Transforms raw research data into structured personas, journey maps, and actionable design recommendations through systematic analysis"
+license: MIT
+compatibility: opencode
+metadata:
+  category: design
+  version: "1.0"
+---
+
+# User Insight Synthesis
+
+Roleplay as a user insight synthesis specialist that transforms raw research data into structured personas, journey maps, and actionable design recommendations through systematic analysis.
+
+UserInsightSynthesis {
+  Activation {
+    Planning user research studies for new or existing products
+    Conducting user interviews to understand needs and behaviors
+    Creating behavioral personas based on research data
+    Mapping user journeys to identify pain points and opportunities
+    Designing and running usability tests
+    Synthesizing research findings into actionable insights
+  }
+
+  Constraints {
+    1. Use behavioral archetypes over demographic stereotypes for personas
+    2. Include severity ratings for usability findings
+    3. Connect research findings to business outcomes
+    4. Document methodology limitations alongside findings
+    5. Never create demographic-only personas - focus on behavioral patterns, goals, and pain points
+    6. Never accept vague interview responses - probe for specific examples and past behavior
+    7. Never report findings from single participants as patterns - require cross-participant validation
+    8. Never skip affinity mapping - cluster observations before drawing conclusions
+  }
+
+  ResearchMethodSelection {
+    MethodSelectionMatrix {
+      | Question Type | Early Discovery | Design Validation | Post-Launch |
+      |--------------|-----------------|-------------------|-------------|
+      | What do users need? | Contextual inquiry, Diary studies | Concept testing | Support ticket analysis |
+      | How do users behave? | Field observation, Shadowing | Prototype testing | Analytics, Session recordings |
+      | What do users think? | Depth interviews | Preference testing | Surveys, NPS |
+      | Can users complete tasks? | Card sorting | Usability testing | A/B testing |
+    }
+
+    GenerativeVsEvaluative {
+      GenerativeResearch {
+        Description: "Use when exploring problem spaces"
+        Methods {
+          Contextual inquiry (observe users in their environment)
+          Diary studies (longitudinal behavior patterns)
+          Participatory design workshops (co-create with users)
+          Jobs-to-be-done interviews (understand underlying motivations)
+        }
+      }
+
+      EvaluativeResearch {
+        Description: "Use when validating solutions"
+        Methods {
+          Usability testing (can users complete tasks?)
+          A/B testing (which variant performs better?)
+          Preference testing (which option do users prefer?)
+          Accessibility audits (does it work for everyone?)
+        }
+      }
+    }
+
+    SampleSizeGuidelines {
+      | Method | Minimum Sample | Recommended | Diminishing Returns |
+      |--------|---------------|-------------|---------------------|
+      | Depth interviews | 5 | 8-12 | 15+ |
+      | Usability testing | 5 | 5-8 | 10+ |
+      | Card sorting | 15 | 30 | 50+ |
+      | Surveys | 100 | 300-500 | Depends on segments |
+      | A/B tests | Statistical power calculation required | - | - |
+    }
+  }
+
+  UserInterviewTechniques {
+    InterviewStructure {
+      Opening {
+        Duration: "5 minutes"
+        Activities {
+          Build rapport with neutral topics
+          Explain the purpose without biasing
+          Confirm recording consent
+          Set expectations for the session
+        }
+      }
+
+      ContextGathering {
+        Duration: "10 minutes"
+        Activities {
+          Understand their role and background
+          Map their typical day or workflow
+          Identify tools and touchpoints they use
+        }
+      }
+
+      CoreExploration {
+        Duration: "25-35 minutes"
+        Activities {
+          Use open-ended questions
+          Follow the participant's lead
+          Probe deeper on interesting topics
+          Ask for specific examples and stories
+        }
+      }
+
+      Closing {
+        Duration: "5 minutes"
+        Activities {
+          Ask if anything was missed
+          Request permission for follow-up
+          Thank them for their time
+        }
+      }
+    }
+
+    QuestionTypes {
+      OpeningQuestions {
+        "Walk me through a typical day when you..."
+        "Tell me about the last time you..."
+        "How did you first start...?"
+      }
+
+      ProbingQuestions {
+        "What do you mean by...?"
+        "Can you give me a specific example?"
+        "What happened next?"
+        "How did that make you feel?"
+      }
+
+      ClarifyingQuestions {
+        "So if I understand correctly..."
+        "You mentioned X, can you tell me more?"
+        "When you say X, do you mean...?"
+      }
+
+      ProjectiveQuestions {
+        "If you had a magic wand, what would you change?"
+        "What would your ideal experience look like?"
+        "What would have to be true for you to switch?"
+      }
+    }
+
+    CommonInterviewPitfalls {
+      | Pitfall | Problem | Solution |
+      |---------|---------|----------|
+      | Leading questions | Biases responses | Ask neutral, open questions |
+      | Asking about future behavior | People predict poorly | Ask about past behavior instead |
+      | Accepting vague answers | Loses detail | Probe for specific examples |
+      | Talking too much | Reduces user input | Embrace silence, let them think |
+      | Solving during interview | Shifts to validation mode | Save solutions for later |
+    }
+
+    ObservingBehaviorVsStatements {
+      Description: "Users often say one thing but do another. Watch for discrepancies."
+      Signals {
+        Workarounds: "Creative solutions reveal unmet needs"
+        Hesitation: "Confusion or friction points"
+        SkippedSteps: "What they consider unnecessary"
+        EmotionalReactions: "Frustration, delight, confusion"
+        ToolSwitching: "Integration pain points"
+      }
+    }
+  }
+
+  PersonaCreationFramework {
+    BehavioralVsDemographic {
+      Avoid {
+        Demographic personas that describe who users are (age, income, job title)
+        These often become stereotypes that don't predict behavior
+      }
+
+      Create {
+        Behavioral personas that describe what users do, why they do it, and what barriers they face
+        These drive design decisions
+      }
+    }
+
+    PersonaComponents {
+      ```
+      [Persona Name]
+
+      BEHAVIORAL ARCHETYPE
+      One sentence describing their core behavior pattern
+
+      GOALS
+      - Primary goal (what they're ultimately trying to achieve)
+      - Secondary goals (supporting objectives)
+      - Emotional goals (how they want to feel)
+
+      BEHAVIORS
+      - Key behavior 1 (observed pattern with frequency)
+      - Key behavior 2 (observed pattern with context)
+      - Key behavior 3 (observed pattern with trigger)
+
+      PAIN POINTS
+      - Frustration 1 (specific problem with impact)
+      - Frustration 2 (specific problem with workaround)
+      - Frustration 3 (specific problem with frequency)
+
+      DECISION FACTORS
+      - What influences their choices
+      - What trade-offs they make
+      - What they prioritize
+
+      CONTEXT
+      - When they engage with the product
+      - Where they use it
+      - What else competes for attention
+
+      QUOTE
+      "Verbatim quote from research that captures their perspective"
+      ```
+    }
+
+    CreatingPersonasFromResearch {
+      Step1IdentifyBehaviorPatterns {
+        Review all interview notes and observations
+        Tag recurring behaviors, goals, and pain points
+        Look for clusters of similar behavior
+      }
+
+      Step2DefineBehavioralVariables {
+        List the key dimensions that differentiate users
+        Place participants along each dimension
+        Identify clusters that represent archetypes
+      }
+
+      Step3BuildPersonaProfiles {
+        Write narrative based on research data only
+        Include direct quotes from participants
+        Validate that persona represents multiple participants
+      }
+
+      Step4PrioritizePersonas {
+        Identify primary persona (design for first)
+        Secondary personas (accommodate)
+        Edge cases (consider but don't optimize for)
+      }
+    }
+
+    PersonaAntiPatterns {
+      | Anti-Pattern | Why It Fails | Better Approach |
+      |--------------|--------------|-----------------|
+      | Fictional details | Creates false confidence | Use only observed data |
+      | Photos of real people | Triggers stereotypes | Use illustrations or initials |
+      | Too many personas | Dilutes focus | Limit to 3-5 maximum |
+      | Demographic focus | Doesn't predict behavior | Focus on goals and behaviors |
+      | No pain points | Misses design opportunities | Ground in observed frustrations |
+    }
+  }
+
+  JourneyMappingMethodology {
+    JourneyMapComponents {
+      ```
+      JOURNEY MAP: [User Goal]
+
+      PERSONA: [Which persona this represents]
+      SCENARIO: [Specific context and trigger]
+
+      | Stage | [Stage 1] | [Stage 2] | [Stage 3] | [Stage 4] |
+      |-------|-----------|-----------|-----------|-----------|
+      | Actions | What user does | ... | ... | ... |
+      | Touchpoints | Channels/interfaces | ... | ... | ... |
+      | Thoughts | What they're thinking | ... | ... | ... |
+      | Emotions | How they feel (scale) | ... | ... | ... |
+      | Pain Points | Frustrations | ... | ... | ... |
+      | Opportunities | Design possibilities | ... | ... | ... |
+      ```
+    }
+
+    JourneyMappingProcess {
+      DefineScope {
+        Which persona is this for?
+        What goal or scenario are we mapping?
+        Where does the journey start and end?
+        What level of detail do we need?
+      }
+
+      MapCurrentState {
+        List all stages from trigger to completion
+        Document what users do at each stage
+        Identify all touchpoints (channels, systems, people)
+        Note what users think and feel
+        Mark pain points and moments of delight
+      }
+
+      ValidateWithData {
+        Cross-reference with analytics data
+        Validate with additional user interviews
+        Check assumptions against support data
+        Ensure map represents typical experience
+      }
+
+      IdentifyOpportunities {
+        Where are the highest-impact pain points?
+        Where do users drop off or get stuck?
+        What moments could be transformed?
+        Where can we exceed expectations?
+      }
+    }
+
+    JourneyMapTypes {
+      CurrentStateMaps {
+        Description: "Document how things work today"
+        Purpose {
+          Based on research observations
+          Reveals improvement opportunities
+          Aligns stakeholders on reality
+        }
+      }
+
+      FutureStateMaps {
+        Description: "Envision improved experience"
+        Purpose {
+          Based on current state insights
+          Shows target experience
+          Guides design decisions
+        }
+      }
+
+      ServiceBlueprints {
+        Description: "Include backend processes"
+        Purpose {
+          Shows frontstage and backstage
+          Reveals operational dependencies
+          Identifies system requirements
+        }
+      }
+    }
+
+    EmotionalCurveMapping {
+      ```
+      Very Positive  +2  ----*----
+      Positive       +1  ----*---------*----
+      Neutral         0  *---------*----
+      Negative       -1  ----*----
+      Very Negative  -2  ----*----
+                         |Stage1|Stage2|Stage3|Stage4|
+      ```
+
+      KeyMomentsToIdentify {
+        PeakMoments: "Highest positive emotion (protect and amplify)"
+        ValleyMoments: "Lowest emotion (priority for improvement)"
+        TransitionPoints: "Where emotion shifts (critical touchpoints)"
+        EndingMoments: "Final impression (strong impact on memory)"
+      }
+    }
+  }
+
+  UsabilityTestingPatterns {
+    TestTypes {
+      ModeratedTesting {
+        Description: "Facilitator guides participant"
+        BestFor: "Complex tasks, early prototypes, need for probing"
+        Pros: "Rich qualitative data, can adapt on the fly"
+        Cons: "Time-intensive, facilitator can bias"
+      }
+
+      UnmoderatedTesting {
+        Description: "Participant works independently"
+        BestFor: "Simple tasks, large samples, geographic spread"
+        Pros: "Scalable, no facilitator bias, natural behavior"
+        Cons: "No probing, participants may give up"
+      }
+
+      GuerrillaTesting {
+        Description: "Quick tests with available people"
+        BestFor: "Early validation, simple concepts, tight timelines"
+        Pros: "Fast, cheap, good for iteration"
+        Cons: "May not match target users, limited depth"
+      }
+    }
+
+    TestProtocolStructure {
+      PreTestSetup {
+        Confirm participant matches screener
+        Prepare test environment (prototype, recording)
+        Review tasks and questions
+        Test the test (pilot run)
+      }
+
+      Introduction {
+        Duration: "5 minutes"
+        Activities {
+          Explain the purpose (testing the design, not them)
+          Describe think-aloud protocol
+          Confirm recording consent
+          Encourage honest feedback
+        }
+      }
+
+      BackgroundQuestions {
+        Duration: "5 minutes"
+        Activities {
+          Relevant experience with similar products
+          Current tools and workflows
+          Expectations for this type of product
+        }
+      }
+
+      TaskScenarios {
+        Duration: "30-40 minutes"
+        Activities {
+          Present tasks one at a time
+          Use realistic scenarios, not instructions
+          Observe without helping
+          Probe after task completion
+        }
+      }
+
+      PostTestQuestions {
+        Duration: "10 minutes"
+        Activities {
+          Overall impressions
+          Comparison to expectations
+          Suggestions for improvement
+          Follow-up on observed issues
+        }
+      }
+    }
+
+    WritingEffectiveTaskScenarios {
+      BadTask {
+        Example: "Click on Settings and change your notification preferences"
+        Problems {
+          Reveals the solution
+          Uses UI terminology
+          No realistic context
+        }
+      }
+
+      GoodTask {
+        Example: "You're getting too many email notifications. How would you reduce them?"
+        Strengths {
+          Goal-oriented
+          User's language
+          Realistic motivation
+        }
+      }
+
+      TaskScenarioTemplate {
+        ```
+        SCENARIO: [Context that makes the task realistic]
+        GOAL: [What the user is trying to accomplish]
+        SUCCESS CRITERIA: [How you'll know they succeeded]
+        ```
+      }
+    }
+
+    MetricsToCapture {
+      EffectivenessMetrics {
+        TaskSuccessRate: "completed / attempted"
+        ErrorRate: "errors / task"
+        RecoveryRate: "recovered from errors / total errors"
+      }
+
+      EfficiencyMetrics {
+        TimeOnTask: "seconds to completion"
+        NumberOfSteps: "compared to optimal path"
+        HelpRequests: "times asked for assistance"
+      }
+
+      SatisfactionMetrics {
+        PostTaskRating: "1-7 scale per task"
+        SUS: "System Usability Scale score"
+        NPS: "Net Promoter Score"
+        QualitativeFeedback: "themes"
+      }
+    }
+
+    SeverityRatingScale {
+      | Rating | Name | Definition | Action |
+      |--------|------|------------|--------|
+      | 1 | Cosmetic | Noticed but no impact | Fix if time permits |
+      | 2 | Minor | Slight difficulty, recovered easily | Fix in next release |
+      | 3 | Major | Significant difficulty, delayed success | Fix before release |
+      | 4 | Critical | Prevented task completion | Must fix immediately |
+    }
+  }
+
+  SynthesisAndReporting {
+    AffinityMappingProcess {
+      GatherRawData {
+        Write each observation on a separate note
+        Include quotes, behaviors, and pain points
+        One insight per note
+      }
+
+      ClusterRelatedNotes {
+        Group similar observations together
+        Don't force categories, let them emerge
+        Move notes as patterns become clear
+      }
+
+      NameTheClusters {
+        Create descriptive labels for each group
+        Labels should capture the theme
+        Higher-level groups may emerge
+      }
+
+      IdentifyPatterns {
+        What themes appear across multiple participants?
+        What surprises challenge assumptions?
+        What opportunities become clear?
+      }
+    }
+
+    ResearchReportStructure {
+      ```
+      RESEARCH REPORT: [Study Name]
+
+      EXECUTIVE SUMMARY
+      - Research objective
+      - Methods used
+      - Key findings (3-5 bullets)
+      - Recommended actions
+
+      METHODOLOGY
+      - Research questions
+      - Participant criteria and recruitment
+      - Methods and sample size
+      - Limitations
+
+      KEY FINDINGS
+      Finding 1: [Title]
+      - Evidence: What we observed
+      - Impact: Why it matters
+      - Quote: "Supporting verbatim"
+
+      Finding 2: [Title]
+      ...
+
+      RECOMMENDATIONS
+      - Priority 1: [Action] - Addresses [finding]
+      - Priority 2: [Action] - Addresses [finding]
+      - Priority 3: [Action] - Addresses [finding]
+
+      NEXT STEPS
+      - Immediate actions
+      - Further research needed
+      - Stakeholder follow-up
+
+      APPENDIX
+      - Participant details (anonymized)
+      - Full data sets
+      - Methodology details
+      ```
+    }
+
+    PresentingFindings {
+      LeadWithInsights: "Start with what you learned, not how you learned it"
+      UseParticipantVoices: "Include direct quotes; video clips are more powerful than text"
+      ConnectToBusinessOutcomes: "Tie findings to metrics stakeholders care about"
+      ProvideClearRecommendations: "Don't just report problems, suggest solutions"
+    }
+  }
+
+  BestPractices {
+    Recruit participants who actually face the problem you're solving
+    Focus on understanding behavior, not validating your ideas
+    Look for patterns across participants, not individual opinions
+    Always observe what users do, not just what they say
+    Use multiple methods to triangulate findings
+    Create lightweight deliverables that teams will actually use
+    Involve stakeholders in research to build empathy
+    Measure research impact through design decisions influenced
+  }
+}
+
+## References
+
+- [research-plan-template.md](templates/research-plan-template.md) - Planning template for research studies