neotoma 0.3.0 → 0.3.2

package/docs/releases/archived/aspirational/v1.0.0/approach_comparison.md

@@ -1,558 +0,0 @@
-# Discovery-First vs. Dogfooding-First: Comparative Analysis
-
-**Date**: 2026-01-15  
-**Purpose**: Assess comparative advantages of discovery-first vs. dogfooding-first approaches for MVP development
-
-## Executive Summary
-
-**Discovery-First Approach:**
-- Discovery (Week -8 to -5) → Build (Week 0-12) → Marketing (Week -4 to Week 4)
-- External validation before building
-- Market-driven feature prioritization
-
-**Dogfooding-First Approach:**
-- Build for personal use (ateles integration) → Release developer preview during dogfooding → Validate through usage → Discovery only if signals are weak
-- Internal validation through real usage
-- Use-case-driven feature prioritization
-
-**Capacity-Constrained Sequence (Recommended When Discovery Is Not Feasible):**
-- Build + dogfood with developer preview release during dogfooding
-- Use preview validation to assess dependence and feedback quality
-- Schedule discovery only if validation is insufficient
-
----
-
-## 1. Time to Value
-
-### Discovery-First
-**Advantages:**
-- ✅ **External validation before investment**: 3-4 weeks of discovery prevents building wrong product
-- ✅ **Parallel activities**: Discovery and marketing prep can run in parallel (Week -8 to -5)
-- ✅ **Clear go/no-go decision**: Discovery findings inform whether to proceed
-
-**Disadvantages:**
-- ❌ **Delayed development start**: 3-4 weeks before building begins
-- ❌ **Discovery overhead**: 40-50 survey responses, 13 interviews, 8 usability tests, 8 pricing interviews
-- ❌ **Total timeline**: ~16-20 weeks (discovery + build + marketing)
-
-**Time Investment:**
-- Discovery: ~80-100 hours (interviews, analysis, synthesis)
-- Marketing prep: ~35 hours (pre-launch) + ~60 hours (post-launch)
-- **Total pre-build**: ~175-195 hours
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Immediate development start**: Begin building immediately
-- ✅ **Faster iteration cycles**: Real usage feedback drives rapid iteration
-- ✅ **No discovery overhead**: Skip 3-4 weeks of discovery activities
-- ✅ **Total timeline**: ~12-14 weeks (build only, marketing deferred)
-
-**Disadvantages:**
-- ❌ **No external validation**: Risk of building for single user (yourself)
-- ❌ **Delayed market feedback**: No external validation until post-dogfooding
-
-**Time Investment:**
-- Discovery: 0 hours (deferred)
-- Marketing: 0 hours (deferred)
-- **Total pre-build**: 0 hours
-
-**Winner**: **Dogfooding-First** (faster time to personal value, but discovery-first has better external validation)
-
----
-
-## 2. Risk Mitigation
-
-### Discovery-First
-**Advantages:**
-- ✅ **Validates problem exists**: 70%+ validation threshold before building
-- ✅ **Validates solution approach**: 60%+ interest in MCP solution before building
-- ✅ **Validates business model**: 50%+ willingness to pay ≥€25/month before building
-- ✅ **Validates technical feasibility**: OCR determinism, MCP reliability tested before building
-- ✅ **Identifies failure modes early**: Pre-mortem failure modes validated/disproven in discovery
-- ✅ **Reduces build risk**: Lower probability of building wrong product
-
-**Disadvantages:**
-- ❌ **Discovery can be wrong**: Survey/interview bias, small sample size (13-29 participants)
-- ❌ **Market can change**: 3-4 weeks + 12 weeks build = 15-16 weeks to market (competitive risk)
-- ❌ **Discovery overhead risk**: Time spent on discovery if product is obvious
-
-**Risk Profile:**
-- **Market fit risk**: Low (validated before building)
-- **Technical risk**: Low (feasibility validated)
-- **Business risk**: Low (pricing validated)
-- **Timing risk**: Medium (15-16 weeks to market)
-- **Competitive risk**: Medium (competitors may launch first)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Validates real usage**: Actual daily usage reveals real problems/needs
-- ✅ **Validates technical feasibility**: Real-world usage tests determinism, performance, reliability
-- ✅ **Validates integration patterns**: Ateles integration validates MCP integration patterns
-- ✅ **Faster failure detection**: Real usage catches issues immediately
-- ✅ **No discovery bias**: Real usage vs. hypothetical interview responses
-
-**Disadvantages:**
-- ❌ **Single-user validation**: Your needs may not represent market needs
-- ❌ **No external validation**: No validation of market demand, pricing, or competitive positioning
-- ❌ **Business model risk**: No validation of willingness to pay
-- ❌ **Market fit risk**: Building for yourself may not solve others' problems
-- ❌ **Competitive positioning risk**: No validation of defensible differentiators
-
-**Risk Profile:**
-- **Market fit risk**: High (single-user validation)
-- **Technical risk**: Low (real usage validates)
-- **Business risk**: High (no pricing validation)
-- **Timing risk**: Low (faster to market)
-- **Competitive risk**: Low (faster to market)
-
-**Winner**: **Discovery-First** (better external risk mitigation), but **Dogfooding-First** has better technical validation through real usage
-
----
-
-## 3. Validation Quality
-
-### Discovery-First
-**Advantages:**
-- ✅ **External validation**: 13-29 participants across multiple segments
-- ✅ **Structured methodology**: Mom Test, Van Westendorp, SUS scores
-- ✅ **Quantitative metrics**: Success criteria with thresholds (70% problem validation, 60% solution interest, 50% willingness to pay)
-- ✅ **Diverse perspectives**: AI-Native Operators + Knowledge Workers
-- ✅ **Commitment signals**: Time/money spent, workarounds built (validates real pain)
-
-**Disadvantages:**
-- ❌ **Hypothetical responses**: Interviews ask about past behavior, but not real usage of Neotoma
-- ❌ **Small sample size**: 13-29 participants may not represent market
-- ❌ **Interview bias**: Participants may be polite, overstate interest
-- ❌ **No real usage data**: Validation based on descriptions, not actual product usage
-
-**Validation Depth:**
-- Problem validation: High (structured interviews, commitment signals)
-- Solution validation: Medium (hypothetical, no real usage)
-- Technical validation: Medium (feasibility tests, but not production usage)
-- Business validation: Medium (pricing interviews, but not real purchases)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Real usage validation**: Actual daily usage reveals real problems, workflows, edge cases
-- ✅ **Integration validation**: Ateles integration validates MCP integration patterns in production
-- ✅ **Technical validation**: Real-world usage tests determinism, performance, reliability at scale
-- ✅ **Feature prioritization**: Real usage drives feature prioritization (what you actually use)
-- ✅ **No hypothetical bias**: Real usage vs. interview responses
-
-**Disadvantages:**
-- ❌ **Single-user validation**: Your needs may not represent market
-- ❌ **No external validation**: No validation of market demand, pricing, competitive positioning
-- ❌ **Limited perspective**: Single user, single use case (ateles integration)
-- ❌ **No quantitative metrics**: Subjective validation vs. structured success criteria
-
-**Validation Depth:**
-- Problem validation: High (real usage, real pain points)
-- Solution validation: High (real usage of actual product)
-- Technical validation: High (production usage, real data, real scale)
-- Business validation: Low (no external validation of pricing, market demand)
-
-**Winner**: **Tie** — Discovery-First has better external/market validation, Dogfooding-First has better technical/usage validation
-
----
-
-## 4. Resource Requirements
-
-### Discovery-First
-**Time Investment:**
-- Discovery: ~80-100 hours (interviews, analysis, synthesis)
-- Marketing prep: ~35 hours (pre-launch) + ~60 hours (post-launch)
-- **Total**: ~175-195 hours before/during build
-
-**Financial Investment:**
-- Discovery incentives: ~$500-1000 (gift cards for participants)
-- Marketing: $0 (organic only)
-- **Total**: ~$500-1000
-
-**Infrastructure:**
-- Discovery tools: Lead sourcing tools, survey platform, interview scheduling
-- Marketing: Social media accounts, website, analytics
-
-### Dogfooding-First
-**Time Investment:**
-- Discovery: 0 hours (deferred)
-- Marketing: 0 hours (deferred)
-- **Total**: 0 hours (all time focused on building)
-
-**Financial Investment:**
-- Discovery: $0
-- Marketing: $0
-- **Total**: $0
-
-**Infrastructure:**
-- Minimal: Just development environment, ateles integration
-
-**Winner**: **Dogfooding-First** (lower resource requirements, faster to start)
-
----
-
-## 5. Market Fit Confidence
-
-### Discovery-First
-**Advantages:**
-- ✅ **External validation**: 13-29 participants validate problem exists
-- ✅ **Market segmentation**: Validates AI-Native Operators vs. Knowledge Workers
-- ✅ **Competitive positioning**: Validates defensible differentiators (privacy, determinism, cross-platform)
-- ✅ **Pricing validation**: Van Westendorp PSM validates price point (€25-€125/month)
-- ✅ **Go/no-go decision**: Clear criteria for proceeding (70% problem validation, 60% solution interest, 50% willingness to pay)
-
-**Confidence Level:**
-- Problem exists: High (70%+ validation threshold)
-- Solution approach: Medium-High (60%+ interest, but hypothetical)
-- Market demand: Medium (13-29 participants, small sample)
-- Pricing: Medium-High (50%+ willingness to pay, but hypothetical)
-- Competitive positioning: Medium (validated in interviews, but not market-tested)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Real usage validation**: Actual daily usage validates product works for real use case
-- ✅ **Integration validation**: Ateles integration validates MCP integration patterns
-- ✅ **Technical validation**: Real-world usage validates determinism, performance, reliability
-
-**Disadvantages:**
-- ❌ **Single-user validation**: Your needs may not represent market
-- ❌ **No external validation**: No validation of market demand, pricing, competitive positioning
-- ❌ **Limited market insight**: Single use case (ateles integration) may not represent broader market
-
-**Confidence Level:**
-- Problem exists: High (you have the problem, real usage)
-- Solution approach: High (real usage validates product works)
-- Market demand: Low (single-user validation, no external validation)
-- Pricing: Low (no pricing validation)
-- Competitive positioning: Low (no external validation of differentiators)
-
-**Winner**: **Discovery-First** (better market fit confidence through external validation)
-
----
-
-## 6. Technical Validation
-
-### Discovery-First
-**Advantages:**
-- ✅ **Feasibility validation**: OCR determinism, MCP reliability tested before building
-- ✅ **Technical risk mitigation**: Identifies technical blockers early
-- ✅ **Architecture validation**: Validates technical approach before implementation
-
-**Disadvantages:**
-- ❌ **Limited scale**: Feasibility tests on 20+ documents, not production scale
-- ❌ **No production usage**: Tests in controlled environment, not real-world usage
-- ❌ **No integration validation**: No validation of MCP integration patterns in production
-
-**Technical Validation Depth:**
-- OCR determinism: High (100+ document tests)
-- MCP reliability: Medium (99.9% uptime target, but not production-tested)
-- Integration patterns: Low (no production integration validation)
-- Performance at scale: Low (no production scale testing)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Production usage**: Real-world usage tests determinism, performance, reliability at scale
-- ✅ **Integration validation**: Ateles integration validates MCP integration patterns in production
-- ✅ **Real data**: Real parquet files, real entity types, real workflows
-- ✅ **Edge case discovery**: Real usage reveals edge cases, bugs, performance issues
-- ✅ **Iteration speed**: Real usage drives rapid iteration on technical issues
-
-**Disadvantages:**
-- ❌ **Single use case**: Ateles integration may not reveal all technical issues
-- ❌ **Limited scale**: Single-user usage may not test scale limits
-
-**Technical Validation Depth:**
-- OCR determinism: High (production usage, real documents)
-- MCP reliability: High (production usage, real queries)
-- Integration patterns: High (production ateles integration)
-- Performance at scale: Medium (single-user, but real data/workloads)
-
-**Winner**: **Dogfooding-First** (better technical validation through production usage)
-
----
-
-## 7. Business Viability
-
-### Discovery-First
-**Advantages:**
-- ✅ **Pricing validation**: Van Westendorp PSM validates price point (€25-€125/month)
-- ✅ **Willingness to pay**: 50%+ threshold validates business model
-- ✅ **Value drivers**: Identifies what drives willingness to pay
-- ✅ **Go/no-go decision**: Clear criteria for business viability
-
-**Disadvantages:**
-- ❌ **Hypothetical pricing**: Interviews ask about willingness to pay, but not real purchases
-- ❌ **Small sample size**: 8 pricing interviews may not represent market
-- ❌ **No revenue validation**: No actual revenue until post-launch
-
-**Business Validation:**
-- Pricing: Medium-High (50%+ willingness to pay, but hypothetical)
-- Business model: Medium (validated in interviews, but not market-tested)
-- Revenue potential: Low (no actual revenue until post-launch)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Real value validation**: Actual daily usage validates product provides value
-- ✅ **Use case validation**: Ateles integration validates specific use case value
-
-**Disadvantages:**
-- ❌ **No pricing validation**: No validation of willingness to pay
-- ❌ **No business model validation**: No validation of subscription model, pricing tiers
-- ❌ **No revenue validation**: No actual revenue until post-dogfooding marketing
-
-**Business Validation:**
-- Pricing: Low (no pricing validation)
-- Business model: Low (no business model validation)
-- Revenue potential: Low (no actual revenue until post-dogfooding)
-
-**Winner**: **Discovery-First** (better business viability validation through pricing interviews)
-
----
-
-## 8. Iteration Speed
-
-### Discovery-First
-**Advantages:**
-- ✅ **Informed iteration**: Discovery findings inform feature prioritization
-- ✅ **Continuous discovery**: Weekly user interviews during development (2-3 participants/week)
-- ✅ **External feedback**: External validation drives iteration
-
-**Disadvantages:**
-- ❌ **Slower iteration**: Discovery → Build → Marketing sequence adds overhead
-- ❌ **Interview overhead**: Weekly interviews add time overhead
-- ❌ **Delayed feedback**: Discovery findings may not match real usage
-
-**Iteration Cycle:**
-- Discovery phase: 3-4 weeks (one-time)
-- Build phase: 12 weeks (with continuous discovery)
-- Marketing phase: 4 weeks (post-launch)
-- **Total**: ~19-20 weeks
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Rapid iteration**: Real usage feedback drives immediate iteration
-- ✅ **No interview overhead**: No discovery interviews, all time on building
-- ✅ **Real-time feedback**: Real usage reveals issues immediately
-- ✅ **Faster cycles**: Build → Use → Iterate cycles are faster
-
-**Disadvantages:**
-- ❌ **Single-user feedback**: Limited to your perspective
-- ❌ **No external feedback**: No external validation until post-dogfooding
-
-**Iteration Cycle:**
-- Build phase: 12 weeks (with real usage feedback)
-- Dogfooding phase: Ongoing (continuous real usage)
-- Marketing phase: Deferred (post-dogfooding)
-- **Total**: ~12 weeks (marketing deferred)
-
-**Winner**: **Dogfooding-First** (faster iteration through real usage feedback)
-
----
-
-## 9. Strategic Positioning
-
-### Discovery-First
-**Advantages:**
-- ✅ **Validates defensible differentiators**: Interviews validate privacy-first, deterministic, cross-platform positioning
-- ✅ **Competitive analysis**: Discovery includes competitive positioning validation
-- ✅ **Market messaging**: Discovery findings inform marketing messaging
-- ✅ **Strategic clarity**: Clear go/no-go decision based on competitive positioning
-
-**Disadvantages:**
-- ❌ **Delayed market entry**: 15-16 weeks to market (competitive risk)
-- ❌ **Competitive timing risk**: Competitors may launch first
-
-**Strategic Validation:**
-- Defensible differentiators: Medium (validated in interviews, but not market-tested)
-- Competitive positioning: Medium (validated in interviews, but not market-tested)
-- Market messaging: Medium (informed by discovery, but not market-tested)
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Faster market entry**: 12 weeks to usable product (competitive advantage)
-- ✅ **Real usage validation**: Real usage validates product works, not just positioning
-- ✅ **Technical differentiation**: Real usage validates technical differentiators (determinism, MCP integration)
-
-**Disadvantages:**
-- ❌ **No external positioning validation**: No validation of defensible differentiators with market
-- ❌ **No competitive analysis**: No validation of competitive positioning
-- ❌ **Limited strategic clarity**: Single-user validation may not represent market
-
-**Strategic Validation:**
-- Defensible differentiators: Low (no external validation)
-- Competitive positioning: Low (no external validation)
-- Market messaging: Low (no external validation)
-
-**Winner**: **Discovery-First** (better strategic positioning validation), but **Dogfooding-First** has faster market entry
-
----
-
-## 10. Use Case Validation (Ateles Integration)
-
-### Discovery-First
-**Advantages:**
-- ✅ **Broad market validation**: Validates problem across multiple segments, not just ateles use case
-- ✅ **Multiple use cases**: Discovery may reveal use cases beyond ateles integration
-
-**Disadvantages:**
-- ❌ **No ateles-specific validation**: Discovery doesn't validate ateles integration use case
-- ❌ **Hypothetical validation**: Interviews ask about use cases, but not real ateles integration
-
-### Dogfooding-First
-**Advantages:**
-- ✅ **Real ateles integration**: Actual integration validates use case works
-- ✅ **Production validation**: Real usage in ateles validates MCP integration patterns
-- ✅ **Parquet replacement validation**: Validates Neotoma can replace parquet MCP in ateles
-- ✅ **Real data validation**: Real parquet files, real entity types, real workflows
-
-**Disadvantages:**
-- ❌ **Single use case**: Ateles integration may not represent broader market
-- ❌ **Limited perspective**: Single use case may miss other important use cases
-
-**Winner**: **Dogfooding-First** (better ateles-specific validation through real integration)
-
----
-
-## 11. Competitive Timing
-
-### Discovery-First
-**Timeline to Market:**
-- Discovery: 3-4 weeks (Week -8 to -5)
-- Build: 12 weeks (Week 0-12)
-- Marketing: 4 weeks (Week -4 to Week 4, parallel with build)
-- **Total**: ~15-16 weeks to market
-
-**Competitive Risk:**
-- Medium: 15-16 weeks gives competitors time to launch first
-- Discovery findings may reveal competitive threats
-
-### Dogfooding-First
-**Timeline to Market:**
-- Build: 12 weeks (immediate start)
-- Dogfooding: Ongoing (parallel with build)
-- Marketing: Deferred (post-dogfooding)
-- **Total**: ~12 weeks to usable product (marketing deferred)
-
-**Competitive Risk:**
-- Low: 12 weeks is faster to market
-- But: Marketing deferred means slower public launch
-
-**Winner**: **Dogfooding-First** (faster to usable product, but marketing deferred)
-
----
-
-## 12. Failure Mode Mitigation
-
-### Discovery-First
-**Pre-Mortem Failure Modes:**
-1. **RLS Implementation Issues**: Discovery doesn't mitigate (technical issue)
-2. **Graph Integrity Regressions**: Discovery doesn't mitigate (technical issue)
-3. **MVP Date Slips**: Discovery adds 3-4 weeks, but reduces scope surprises
-4. **OCR Determinism Fails**: ✅ **Discovery mitigates** (feasibility validation)
-5. **Discovery Reveals Low Willingness-to-Pay**: ✅ **Discovery mitigates** (go/no-go decision)
-
-**Mitigation Effectiveness:**
-- Technical risks: Low (discovery doesn't mitigate technical issues)
-- Market risks: High (discovery mitigates market fit, pricing risks)
-- Business risks: High (discovery mitigates business viability risks)
-
-### Dogfooding-First
-**Pre-Mortem Failure Modes:**
-1. **RLS Implementation Issues**: ✅ **Dogfooding mitigates** (real usage catches issues)
-2. **Graph Integrity Regressions**: ✅ **Dogfooding mitigates** (real usage catches issues)
-3. **MVP Date Slips**: ✅ **Dogfooding mitigates** (faster iteration, no discovery overhead)
-4. **OCR Determinism Fails**: ✅ **Dogfooding mitigates** (real usage catches issues)
-5. **Discovery Reveals Low Willingness-to-Pay**: ❌ **Dogfooding doesn't mitigate** (no pricing validation)
-
-**Mitigation Effectiveness:**
-- Technical risks: High (dogfooding mitigates technical issues through real usage)
-- Market risks: Low (dogfooding doesn't mitigate market fit, pricing risks)
-- Business risks: Low (dogfooding doesn't mitigate business viability risks)
-
-**Winner**: **Tie** — Discovery-First mitigates market/business risks, Dogfooding-First mitigates technical risks
-
----
-
-## 13. Recommended Approach by Scenario
-
-### Choose Discovery-First If:
-- ✅ **Market validation is critical**: Need external validation of problem, solution, pricing
-- ✅ **Business viability is uncertain**: Need validation of willingness to pay, business model
-- ✅ **Competitive positioning is critical**: Need validation of defensible differentiators
-- ✅ **Resources available**: Have time/budget for discovery activities
-- ✅ **Market timing is flexible**: 15-16 weeks to market is acceptable
-
-### Choose Dogfooding-First If:
-- ✅ **Personal use case is clear**: Ateles integration is primary use case
-- ✅ **Technical validation is critical**: Need real-world usage to validate determinism, MCP integration
-- ✅ **Speed to value is critical**: Need usable product quickly (12 weeks vs. 15-16 weeks)
-- ✅ **Resources are limited**: Want to focus all time on building
-- ✅ **Market validation can be deferred**: Can validate market fit after dogfooding
-
----
-
-## 14. Hybrid Approach (Recommended)
-
-**Best of Both Worlds:**
-1. **Start with dogfooding**: Build MVP for ateles integration (12 weeks)
-2. **Validate through usage**: Real usage validates technical feasibility, integration patterns
-3. **Then conduct discovery**: After dogfooding, conduct discovery with working product
-4. **Iterate based on findings**: Discovery findings inform v1.1+ features
-5. **Market after validation**: Launch marketing after both dogfooding and discovery validation
-
-**Advantages:**
-- ✅ Faster time to personal value (12 weeks)
-- ✅ Real usage validates technical feasibility
-- ✅ Discovery validates market fit with working product (stronger validation)
-- ✅ Lower risk: Technical validation first, then market validation
-- ✅ Better discovery: Can show working product in discovery interviews
-
-**Timeline:**
-- Build + Dogfooding: 12 weeks
-- Discovery: 3-4 weeks (with working product)
-- Marketing: 4 weeks (post-discovery)
-- **Total**: ~19-20 weeks (same as discovery-first, but with working product earlier)
-
----
-
-## 15. Conclusion
-
-**Discovery-First Advantages:**
-- Better external/market validation
-- Better business viability validation
-- Better strategic positioning validation
-- Lower market fit risk
-- Lower business model risk
-
-**Dogfooding-First Advantages:**
-- Faster time to personal value
-- Better technical validation through real usage
-- Better ateles-specific validation
-- Lower resource requirements
-- Faster iteration cycles
-- Lower technical risk
-
-**Recommendation:**
-- **For ateles integration use case**: **Dogfooding-First** is recommended
-  - Clear use case (ateles integration)
-  - Technical validation is critical (determinism, MCP integration)
-  - Speed to value is important
-  - Market validation can be deferred
-
-- **For broader market launch**: **Hybrid Approach** is recommended
-  - Start with dogfooding (12 weeks)
-  - Then conduct discovery with working product (3-4 weeks)
-  - Then launch marketing (4 weeks)
-  - Total: ~19-20 weeks, but with working product earlier
-
-**Key Insight:**
-Dogfooding-first is optimal when:
-1. Personal use case is clear and representative
-2. Technical validation is critical
-3. Speed to value is important
-4. Market validation can be deferred
-
-Discovery-first is optimal when:
-1. Market validation is critical
-2. Business viability is uncertain
-3. Competitive positioning is critical
-4. Resources are available for discovery