analyze

SaaS Analysis Pipeline — Orchestrator

You are running the full SaaS validation pipeline for: $ARGUMENTS

Setup

1. Create a URL-safe slug from the product name (lowercase, hyphens, no special chars)
2. Create directory: output/<slug>/
3. Save metadata: output/<slug>/.meta.json with {"input": "$ARGUMENTS", "slug": "<slug>", "started": "<ISO date>"}

Pipeline

Execute these phases sequentially. Each phase is interactive — present findings and get user feedback before moving to the next.

Phase 1: Discovery

Run the discovery research for $ARGUMENTS:

• Search the web for the product's website, documentation, pricing, blog, press coverage
• Fetch and analyze key pages
• Write findings to output/<slug>/01-discovery.md with sections:
  1. Product Overview — name, URL, tagline, founding year, team size, funding
  2. Core Features — detailed list organized by category
  3. Pricing Model — tiers, limits, feature gates, free tier
  4. Target Market — ICP, segments, use cases
  5. Technology Stack — from job postings, docs, GitHub, BuiltWith
  6. Traction Signals — traffic estimates, review counts, G2/Capterra ratings, social mentions
  7. Business Model Analysis — revenue model, unit economics indicators

After writing: Present a summary of key findings. Ask the user:

• "Is this accurate? Anything to add or correct?"
• "Any specific areas to focus on for competitive analysis?"

Incorporate feedback before proceeding.

Phase 2: Competitive Analysis

• Read output/<slug>/01-discovery.md for context
• Search for 5-10 direct competitors and 3-5 indirect/adjacent competitors
• Use the Agent tool to spawn parallel researcher subagents — one per competitor — for faster research
• Write findings to output/<slug>/02-competitive.md with sections:
  1. Competitive Landscape Map — categorized list of all competitors
  2. Feature Comparison Matrix — markdown table comparing features
  3. Pricing Comparison — tier-by-tier comparison
  4. Differentiation Opportunities — gaps, underserved segments
  5. Competitive Threats — moats, network effects, switching costs
  6. Market Size Estimation — TAM/SAM/SOM with sources

After writing: Present the competitor matrix. Ask the user:

• "Are there competitors I missed? Any I should research deeper?"
• "Which differentiation opportunities look most promising to you?"

If user names additional competitors, research them and update the matrix.

Phase 3: Business Validation

• Read output/<slug>/01-discovery.md and output/<slug>/02-competitive.md
• Present analysis section by section, asking for user feedback at each:

Use AskUserQuestion for each major section:

1. Problem-Solution Fit — Present assessment → ask "Do you agree?"
2. Value Proposition Canvas (JTBD, pains, gains) → ask "What resonates? What's missing?"
3. SWOT Analysis → ask "What would you change or add?"
4. Risk Assessment Matrix (rated High/Med/Low) → ask "How do you see these risks?"
5. Go-to-Market Feasibility → ask "What channels do you have access to?"
6. MVP Scope Recommendation → ask "What are YOUR priorities?"
7. Verdict: GO / CONDITIONAL GO / NO-GO → ask "Your thoughts?"

Iterate until user explicitly approves the verdict. Write final version to output/<slug>/03-validation.md.

Phase 4: Architecture Discussion

• Read all prior outputs
• This phase mimics Plan mode — present options with trade-offs, let user choose

Use AskUserQuestion with options and previews for each decision:

1. Tech Stack — 2-3 options with pros/cons/cost → user picks
2. Architecture Style — monolith vs microservices vs serverless → user picks
3. Data Model — present mermaid ER diagram draft → user refines
4. API Design — present endpoint structure → user adjusts
5. Frontend Approach — SPA vs SSR vs hybrid → user picks
6. Infrastructure — hosting options with cost estimates → user picks

Record each decision in output/<slug>/04-architecture-decisions.md.

Phase 5: Implementation Spec

• Read all prior outputs, especially architecture decisions
• Generate comprehensive output/<slug>/05-implementation.md with:
  1. System Architecture (mermaid diagram)
  2. Chosen Tech Stack with reasoning
  3. Data Model (mermaid ER diagram)
  4. API Design (endpoints, auth strategy)
  5. Feature Breakdown (epics, user stories, P0/P1/P2)
  6. Implementation Phases (MVP → Phase 1 → Phase 2)
  7. Infrastructure (deploy, CI/CD, monitoring, cost estimate)
  8. Third-Party Integrations
  9. Security Considerations
  10. Success Metrics (KPIs)

After writing: Present summary. Ask:

• "Any changes to feature priorities?"
• "Is the implementation timeline realistic?"

Incorporate feedback.

Phase 6: Ralph Loop Prompt

• Read the spec
• Generate output/<slug>/RALPH-PROMPT.md containing a Ralph Loop command:
  • Clear implementation steps ordered from the spec
  • Checkable success criteria derived from P0 features
  • Completion promise: "DONE"
  • Suggested --max-iterations based on project scope

Present the draft to user. Ask:

• "Adjust success criteria? Change max iterations? Any constraints?"

Generate final command after user approval.

Final: Bundle

• Merge phases 1-5 into output/<slug>/SPEC.md as a single self-contained document
• Print summary: verdict, key architecture decisions, and the Ralph Loop command to run