User research

User Research

Understand users to build products they love.

What This Skill Does

Conducts user research through various methods including interviews, surveys, persona creation, and user behavior analysis. The output provides deep user insights that drive product decisions and feature prioritization.

When to Use

Activate this skill when:

• User asks about understanding user needs or behaviors
• Phrases like "user research", "talk to users", "what do users want"
• Creating or updating user personas
• Analyzing user feedback or interview data
• Need to validate product ideas with real users
• User says "who is our target user" or "create personas"

How It Works

The research process ensures comprehensive user understanding:

1. Define research goals - Identify what you need to learn
2. Select methods - Choose appropriate research methods
3. Create instruments - Design interview guides, surveys
4. Gather data - Conduct research sessions
5. Analyze findings - Identify patterns and insights
6. Create artifacts - Generate personas, journey maps, insights

Input Parameters

Parameter	Type	Required	Description
research_goal	string	Yes	What you want to learn about users
target_segment	string	No	Specific user segment to study
method	string	No	interviews, surveys, personas, all (default)
sample_size	number	No	Target number of participants

Research Methods

Method	When to Use	Output
interviews	Deep understanding of needs	Interview transcripts, key quotes
surveys	Broad pattern validation	Survey results, statistical insights
personas	User archetype creation	Persona profiles
journey-mapping	Experience optimization	User journey maps

Anti-Hallucination Rules (Self-Contained)

The following rules are mandatory for this skill. They are inlined here for standalone installation compatibility.

1. Mandatory Search First

引用行业报告、统计数据前必须先搜索验证。

• 禁止仅凭训练记忆输出任何具体数据（用户数据、统计趋势等）
• 每个数据点必须来自实际搜索和访问的页面
• 搜索查询应包含年份，例如 user research trends 2025 2026

2. Every Claim Must Have a Source

每个行业数据、统计趋势、用户引用必须有来源。

• 没有来源 = 不得写入
• 来源格式：{ "claim": "...", "source": "https://...", "source_name": "...", "fetched_at": "..." }

3. Unknown is Acceptable

无法确认的数据标注 "Unknown"，禁止编造用户画像数据。标注未知是诚实，不是失败。

4. Confidence Rating

关键数据点标注 confidence (high/medium/low)。

• high: 权威研究报告、官方统计数据
• medium: 可靠媒体、行业报告、公开调查
• low: 社区讨论、博客、推测性分析

5. Distinguish Fact from Inference

区分真实用户反馈和推断的用户需求。推断必须标注 basis 和 confidence。

6. Quality Gate (Non-negotiable)

以下任一条未满足，不得输出结果：

• 所有具体数据/统计声明都有来源 URL
• 无法验证的数据已标注 "Unknown"
• 每个数据点标注了 confidence 等级
• facts 和 inferences 已明确区分
• 搜索记录已附在输出中

Search Record

每个分析输出必须附带搜索记录：{ "search_queries_used": [...], "sources_accessed": [{ "url": "...", "title": "...", "used_for": "..." }] }

Output Structure

The skill generates two outputs:

1. JSON file (docs/product/.ompm/user-research.json) - Structured research data
2. Markdown report - Human-readable research summary

JSON Output Format

{
  "research": {
    "id": "uuid",
    "timestamp": "2026-03-12T...",
    "goal": "Understand onboarding pain points",
    "method": "interviews",
    "participants": 8,
    "personas": [
      {
        "name": "Product Manager Alice",
        "role": "PM at B2B SaaS",
        "goals": ["Ship features faster", "Align team"],
        "pain_points": ["Too many tools", "Unclear priorities"],
        "behaviors": ["Checks Jira first thing", "Prefers async comms"],
        "confidence": "medium",
        "source": { "type": "interview", "url_or_reference": "interview-2026-03-12-01", "fetched_at": "2026-03-12T..." }
      }
    ],
    "key_insights": [
      { "insight": "Users spend 40% of time context-switching between tools", "confidence": "high", "evidence_count": 6, "total_participants": 8 }
    ],
    "quotes": [
      {"user": "PM Alice", "quote": "I just wish everything was in one place", "source": "interview-2026-03-12-01"}
    ],
    "recommendations": [
      "Add integrations with popular tools",
      "Simplify onboarding flow"
    ]
  },
  "search_record": {
    "search_queries_used": ["B2B SaaS PM user pain points study 2025"],
    "sources_accessed": [{ "url": "https://...", "title": "...", "used_for": "industry benchmark data" }]
  },
  "last_updated": "2026-03-12T..."
}

Markdown Report Structure

# User Research Report

## Research Overview
- **Goal**: Understand onboarding pain points
- **Method**: User Interviews
- **Participants**: 8 users
- **Date**: YYYY-MM-DD

## User Personas

### Product Manager Alice
**Role**: PM at B2B SaaS company (50-200 employees)

**Goals**:
- Ship features faster without breaking things
- Keep team aligned on priorities
- Get visibility into what engineering is working on

**Pain Points**:
- Too many tools to check (Jira, Slack, GitHub, Notion)
- Unclear priorities lead to context-switching
- Hard to track progress without micromanaging

**Behaviors**:
- Checks Jira first thing every morning
- Prefers async communication over meetings
- Creates detailed specs but engineers don't read them

**Quote**: *"I just wish everything was in one place instead of checking 5 different apps."*

### Engineer Bob
**Role**: Senior developer at startup

**Goals**:
- Focus on coding, not process overhead
- Understand what to work on next
- Get unblocked quickly when stuck

**Pain Points**:
- Requirement ambiguity leads to rework
- Waiting for PM clarification blocks progress
- Too many meetings interrupt flow

**Quote**: *"By the time I get clarification, I've lost my flow state."*

## Key Insights

### Insight 1: Context-Switching Friction
**Finding**: Users spend 40% of time switching between tools
**Evidence**: 6/8 participants mentioned this as top frustration
**Opportunity**: Unified workspace could save hours daily

### Insight 2: Onboarding Drop-off
**Finding**: Users abandon setup at team invitation step
**Evidence**: 50% of interviews cited setup complexity
**Opportunity**: Simplify or defer team setup

### Insight 3: Async Preference
**Finding**: Knowledge workers prefer async updates
**Evidence**: 7/8 prefer written over meetings
**Opportunity**: Build async status updates

## Recommendations

1. **Short-term**: Add status page showing all work in one view
2. **Medium-term**: Build integrations with Jira, GitHub, Slack
3. **Long-term**: Create AI-powered priority recommendations

## Next Steps
- [ ] Validate persona hypotheses with more users
- [ ] Run A/B test on simplified onboarding
- [ ] Prototype unified status view

Persona Template

### [Name]
**Role**: [Job title] at [Company type]

**Goals**:
- [Primary goal]
- [Secondary goal]

**Pain Points**:
- [Key frustration 1]
- [Key frustration 2]

**Behaviors**:
- [Observable behavior 1]
- [Observable behavior 2]

**Quote**: *"[Direct user quote]"*

**Demographics**:
- Company size: [X-Y employees]
- Tech stack: [Key tools]
- Experience level: [Years]

Interview Guide Template

## Interview Questions

### Warm-up
1. Tell me about your role
2. What does a typical day look like?

### Discovery
3. How do you currently [solve problem X]?
4. What tools do you use?
5. What works well? What doesn't?

### Deep Dive
6. Walk me through the last time you [encountered problem]
7. What was frustrating about that?
8. What did you try instead?

### Closing
9. If you could change one thing, what would it be?
10. Is there anything else I should know?

Quality Standards

Before delivering, the research should:

• Include 3+ personas (if applicable) — base on real data, do NOT fabricate
• Provide real insights with evidence counts from actual research
• Include direct user quotes with source references
• ALL industry benchmarks and statistics have source URLs
• ALL quantitative claims have confidence rating
• Facts (real interview data) and inferences (assumed needs) are clearly distinguished
• Search record is included proving searches were executed for industry data
• Link insights to actionable recommendations
• Be valid JSON for downstream skills

Context Integration

Reads:

• docs/product/.ompm/market-analysis.json - Market context for user segments

Writes:

• docs/product/.ompm/user-research.json - Research results for other skills

Read By:

• product-positioning - Uses personas for positioning strategy
• prd-gen - References user needs in requirements
• prioritization - Uses insights for feature prioritization

Example Usage

User: "Create personas for our product"
→ Activates user-research with method=personas

User: "What are our users' biggest pain points?"
→ Activates user-research with research_goal="identify pain points"

User: "Analyze these user interviews"
→ Activates user-research to analyze provided interview data