## Research Framework ### Phase 1: Research Scoping ``` I need to conduct research on: **Topic:** [Research question or subject] **Context:** [Why this research is needed, stakeholders] **Scope:** [What's in scope vs. out of scope] **Deliverable:** [Report, presentation, decision memo] **Timeline:** [When is this due] Help me define a comprehensive research plan: 1. **Research Questions** - Primary question: [Main question to answer] - Sub-questions: [Supporting questions] - Success criteria: [What constitutes a complete answer] 2. **Information Sources** - Academic: [Journals, papers, books] - Industry: [Reports, whitepapers, trade publications] - Primary data: [Interviews, surveys, experiments] - Secondary data: [Databases, public records, APIs] - Web sources: [News, blogs, forums] 3. **Research Methodology** - Qualitative vs. quantitative approach - Sample size and selection criteria (if applicable) - Data collection methods - Analysis frameworks to apply 4. **Potential Biases & Limitations** - Known blind spots in available data - Conflicts of interest in sources - Recency of information (outdated data?) - Geographic or demographic coverage gaps 5. **Outline & Structure** - Table of contents for final report - Logical flow of argument - Key sections and subsections Provide a detailed research plan and timeline. ``` ### Phase 2: Information Gathering #### Systematic Literature Review ``` Conduct a systematic literature review on: **Topic:** [Specific research area] **Databases:** [PubMed, Google Scholar, JSTOR, arXiv, etc.] **Timeframe:** [e.g., last 10 years] **Inclusion Criteria:** [What qualifies a source for inclusion] **Exclusion Criteria:** [What to filter out] Search Strategy: 1. **Keyword Generation** - Primary keywords: [Main terms] - Synonyms and related terms: [Variations] - Boolean operators: [AND, OR, NOT combinations] 2. **Source Identification** - Run searches across databases - Check citation networks (who cited whom) - Review bibliographies of key papers - Identify seminal works and recent developments 3. **Screening & Selection** - Title/abstract screening for relevance - Full-text review for included papers - Quality assessment (methodology, sample size, peer review) - Extract key findings into structured format 4. **Data Extraction** For each source, extract: - Author, year, publication - Research question and methodology - Key findings and conclusions - Limitations and future research needs - Relevance to our research question 5. **Synthesis** - Identify themes and patterns across sources - Note areas of consensus vs. disagreement - Highlight gaps in existing research - Map evolution of thinking over time Generate a literature review matrix and narrative summary. ``` #### Competitive Intelligence Research ``` Conduct competitive intelligence on: **Company/Product:** [Target to research] **Purpose:** [Market entry, M&A, partnership, etc.] **Competitive Set:** [Direct and indirect competitors] Research Plan: 1. **Company Overview** - Founding, history, major milestones - Leadership team and key personnel - Corporate structure and ownership - Financials (if public): revenue, growth, profitability 2. **Product/Service Analysis** - Features and capabilities - Pricing and packaging - Target customers and segments - Strengths and weaknesses (SWOT) 3. **Market Position** - Market share and growth trends - Customer reviews and sentiment - Brand perception and reputation - Win/loss analysis (why customers choose them) 4. **Go-to-Market Strategy** - Sales channels and partnerships - Marketing approach and messaging - Geographic presence - Recent campaigns and initiatives 5. **Technology Stack** - Tech infrastructure (inferred from job postings, stack overflow, etc.) - Patent portfolio and IP - R&D investment and innovation pipeline 6. **Strategic Direction** - Recent acquisitions or divestitures - Executive statements and earnings calls - Job postings (what they're hiring for) - Industry analyst commentary Deliverable: Competitive profile with recommendations. ``` #### Expert Synthesis from Multiple Documents ``` I have multiple sources on the same topic: **Sources:** 1. [Document 1: type, summary] 2. [Document 2: type, summary] 3. [Document 3: type, summary] ... Please synthesize these into a unified analysis: 1. **Compare & Contrast** - What do all sources agree on? - Where do they disagree or contradict? - How do you explain the discrepancies? 2. **Triangulation** - Which claims are supported by multiple independent sources? - Which claims appear in only one source (less reliable)? - Rate confidence level for each finding 3. **Timeline Construction** - If sources cover a time period, create a chronological narrative - Identify cause-and-effect relationships - Note inflection points or turning points 4. **Hierarchy of Information** - What are the most important findings? (top 3-5) - What are supporting details? - What can be relegated to appendices? 5. **Knowledge Gaps** - What questions remain unanswered? - What additional research would be most valuable? Provide integrated narrative with citations. ``` ### Phase 3: Analysis & Insights #### Root Cause Analysis ``` Based on research findings, conduct root cause analysis for: **Problem/Phenomenon:** [What we're investigating] **Observable Symptoms:** [What we can see happening] **Hypotheses:** [Potential explanations] **5 Whys Analysis:** - Why does [problem] occur? [Answer] - Why is that? [Deeper answer] - Why is that? [Even deeper] - Why is that? - Why is that? [Root cause] **Fishbone Diagram:** Organize potential causes by category: - People: [Human factors] - Process: [Workflow issues] - Technology: [System problems] - Environment: [External factors] - Policy: [Regulatory or organizational] **Evidence Mapping:** For each potential cause: - Evidence supporting: [Data, sources] - Evidence against: [Contradictory data] - Confidence level: [High/Medium/Low] **Conclusion:** Most likely root cause(s) with supporting evidence. ``` #### Trend Analysis & Forecasting ``` Analyze trends from research data: **Data:** [Describe data sources and time series] **Objective:** [What we want to forecast or understand] Analysis Steps: 1. **Historical Pattern Recognition** - Identify cyclical patterns (seasonal, economic cycles) - Detect long-term trends (growth, decline) - Note anomalies or outliers 2. **Driver Identification** - What factors influence this trend? - Correlate with external variables (GDP, tech adoption, etc.) - Identify leading vs. lagging indicators 3. **Scenario Modeling** - Base case: Current trajectory continues - Bull case: Positive developments accelerate trend - Bear case: Negative developments reverse trend - Black swan: Unexpected disruptive event 4. **Forecast Confidence** - How reliable is historical data? - How stable are underlying drivers? - What could invalidate the forecast? 5. **Implications** - What should we do if forecast is accurate? - What are early warning signs if forecast is wrong? - Contingency planning for different scenarios Generate forecast report with charts and confidence intervals. ``` #### Framework Application ``` Apply analytical framework to research findings: **Framework:** [SWOT, Porter's Five Forces, PESTLE, BCG Matrix, etc.] **Context:** [Industry, company, situation being analyzed] **SWOT Example:** **Strengths:** - [Internal positive factors from research] - [Competitive advantages identified] **Weaknesses:** - [Internal limitations found] - [Areas where competitors outperform] **Opportunities:** - [External favorable trends] - [Market gaps or unmet needs] **Threats:** - [External risks or challenges] - [Competitive or regulatory headwinds] **Strategic Recommendations:** - Leverage strengths to capitalize on opportunities - Address weaknesses before threats materialize - Prioritized action plan with rationale Apply framework systematically with evidence citations. ``` ### Phase 4: Report Generation #### Executive Summary Creation ``` Based on the research conducted, create an executive summary: **Audience:** [Decision-makers, stakeholders] **Length:** [1-2 pages max] **Tone:** [Formal, action-oriented] **Structure:** 1. **Research Question** (2-3 sentences) - What we sought to understand - Why it matters 2. **Methodology** (1 paragraph) - How research was conducted - Sources and sample size - Timeframe 3. **Key Findings** (3-5 bullet points) - Most important discoveries - Quantify where possible - Highlight surprises 4. **Insights & Implications** (1-2 paragraphs) - What findings mean for the organization - Strategic or operational implications - Risks and opportunities 5. **Recommendations** (3-5 actions) - Specific, actionable next steps - Prioritized by impact and feasibility - Owners and timelines Format for readability: short paragraphs, bullet points, bold key terms. ``` #### Comprehensive Research Report ``` Generate full research report: **Table of Contents:** 1. Executive Summary 2. Introduction & Background 3. Methodology 4. Findings 5. Analysis 6. Conclusions 7. Recommendations 8. Appendices **Formatting Guidelines:** - Headers and subheaders for navigation - Charts, graphs, tables for data visualization - Callout boxes for key insights - Footnotes for citations (APA, MLA, Chicago style) - Glossary for technical terms **Narrative Flow:** - Introduce each section with context - Present findings logically (not chronologically by research date) - Use signposting: "First..., Second..., Finally..." - Transitions between sections for coherence **Evidence Standards:** - Every claim backed by citation - Distinguish facts from opinions/interpretations - Note confidence levels for uncertain findings - Include minority views or dissenting evidence Generate draft report with placeholders for visuals. ``` ## Advanced Techniques ### Multi-Disciplinary Synthesis ``` When research spans multiple fields (e.g., technology + law + economics): 1. **Domain Expert Perspective** - Summarize findings from each discipline's viewpoint - Identify discipline-specific assumptions or blind spots 2. **Concept Translation** - Explain technical concepts in layman's terms - Bridge jargon across fields 3. **Integration** - How do insights from different disciplines interact? - Are there conflicts or contradictions to resolve? - Synergies or emergent insights? 4. **Holistic Recommendation** - Recommendations that consider all perspectives - Trade-offs between competing priorities (e.g., innovation vs. regulation) Use this for complex policy research, technology assessments, or market analysis. ``` ### Longitudinal Research Tracking ``` For ongoing research (quarterly competitive updates, market tracking): **Baseline Establishment:** - Document current state comprehensively - Identify metrics to track over time **Change Detection:** - At each update, compare to baseline - Highlight what's new or different - Track velocity of change (accelerating vs. stable) **Cumulative Insights:** - Maintain running document with updates - Build knowledge graph of relationships - Identify long-term trends not visible in point-in-time analysis **Alerting:** - Define thresholds for significant changes - Trigger deeper investigation when crossed ``` ## Best Practices 1. **Source Diversity**: Don't rely on single source type; triangulate 2. **Recency Check**: Always note date of information; flag outdated data 3. **Bias Awareness**: Consider author motivations and potential conflicts 4. **Iterative Refinement**: Start broad, narrow focus as patterns emerge 5. **Version Control**: Track report drafts and major revisions 6. **Peer Review**: Have research validated by subject matter expert if possible 7. **Citation Discipline**: Cite as you research, not after 8. **Archive Sources**: Save copies of web sources (they may disappear) ## Example Session **User**: "Research the market for AI coding assistants and identify opportunities for a new entrant." **Opus Response**: > I'll conduct comprehensive market research: > > **Phase 1: Market Sizing** > Based on industry reports (Gartner, IDC): > > - TAM: $5.2B in 2026, growing to $12.8B by 2030 (25% CAGR) > - Current adoption: 30% of developers use AI coding tools > - Enterprise adoption lags SMB (18% vs. 38%) > > **Phase 2: Competitive Landscape** > Main players identified: > > 1. **GitHub Copilot** (market leader, ~40% share) > - Strengths: GitHub integration, OpenAI tech > - Weaknesses: Limited to VSCode/Neovim > - Pricing: $10-19/user/month > 2. **Cursor** (fast-growing challenger, ~15% share) > - Strengths: Multi-model, composer mode > - Weaknesses: No enterprise features > - Pricing: $20/month > 3. **Tabnine** (privacy-focused, ~10% share) > - Strengths: On-prem deployment > - Weaknesses: Less capable models > - Pricing: $12-39/month > > **Phase 3: Opportunity Analysis** > Based on customer reviews (G2, Reddit, dev forums): > > **Underserved Needs:** > > 1. **Team collaboration** (89% of mentions): Developers want to share prompts, review AI-generated code together > 2. **Domain-specific models** (67%): Finance, healthcare need specialized knowledge > 3. **Code quality enforcement** (54%): AI should enforce team's style guide, not generic best practices > > **Market Gaps:** > > - No strong player in regulated industries (healthcare, finance) > - Limited multi-IDE support (most focus on VSCode) > - Weak test generation capabilities > > **Recommended Positioning:** > "Enterprise-grade AI coding assistant with team collaboration, compliance, and domain expertise" > > Shall I generate the full competitive analysis and GTM strategy? ## Related Resources - [Connected Papers](https://www.connectedpapers.com/) for academic research mapping - [Crunchbase](https://www.crunchbase.com/) for company intelligence - [Statista](https://www.statista.com/) for market data - [Zotero](https://www.zotero.org/) for citation management