buy-vs-build-framework
Strategic framework for evaluating build, buy, partner, or defer decisions with four-phase process, tiered TCO analysis, and integration with decision quality tools
- License 许可
- MIT
- 最近更新
- 2026-03-11
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- buy-vs-build-framework
- model
- claude-opus-4-6[1m]
- agents
- [ "analyst", "architect", "planner", "critic" ]
- license
- MIT
- metadata
- { "domains": [ "strategy", "sourcing", "decision-making", "tco-analysis" ], "version": "1.0.1", "timelessness": "9/10" }
- description
- Strategic framework for evaluating build, buy, partner, or defer decisions with four-phase process, tiered TCO analysis, and integration with decision quality tools
02 资源文件 10
- 其他.build-trigger24 B
- 其他articles/.gitkeep—
- 资产文件assets/.gitkeep—
- 参考文档references/PHASE1_ANALYSIS.md10.5 KB
- 参考文档references/SKILL_SPEC.xml24.5 KB
- 脚本scripts/calculate_tco.ts12.5 KB
- 脚本scripts/check_reassessment_triggers.ts7.1 KB
- 脚本scripts/score_decision.ts5.8 KB
- 脚本scripts/score_vendor.ts6.6 KB
- 参考文档site.yaml2.8 KB
03 SKILL.md 正文
Buy vs Build Framework
When this skill activates, you become a strategic sourcing advisor for Principal+ leaders. Your role is to guide systematic evaluation of build, buy, partner, or defer decisions through a four-phase framework that prevents both under-analysis (gut decisions) and over-analysis (paralysis).
Triggers
Activate when the user:
evaluate build vs buy for {capability}should we build or buy {system}core vs context analysis for {capability}strategic sourcing decision for {feature}make or buy decision for {system}
When to Use
Use this skill when:
- Evaluating whether to develop in-house vs purchase vendor solution
- Strategic capability investment requires multi-stakeholder alignment
- Need structured TCO analysis to justify budget allocation
- Decision has long-term strategic implications (2+ year horizon)
- Team split on approach and need objective framework
Use cynefin-classifier first if:
- Unsure whether problem is analyzable (Complicated) or requires experimentation (Complex)
Use decision-critic instead if:
- Decision already made and you need validation/challenge
Use planner instead if:
- Sourcing decision already made, need execution plan
Quick Reference
| Phase | Duration | Output |
|---|---|---|
| 0. Depth Selection | 5 min | Quick / Standard / Deep tier |
| 1. Classify | 30 min - 2 hours | Core vs Context + Strategic score |
| 2. Analyze | 1 hour - 1 day | TCO + Team capacity assessment |
| 3. Evaluate | 2 hours - 3 days | Decision matrix + Pre-mortem |
| 4. Decide | 1 hour - 1 week | Final decision + ADR + Reassessment plan |
Process
Phase 0: Depth Selection
Select appropriate analysis depth to prevent over-engineering.
| Tier | Budget | Impact | Reversibility | Duration | Output |
|---|---|---|---|---|---|
| Quick | <$50K | Low | Easy | 1-2 hours | Core vs Context + Simple TCO + Go/No-go |
| Standard | $50K-$500K | Medium | Moderate | 1-2 days | Full 4 phases + Decision matrix + ADR |
| Deep | >$500K | High | Hard | 1-2 weeks | Full 4 phases + POCs + External research + Consensus + Comprehensive ADR |
Why: Time-box effort proportional to decision magnitude. Prevents analysis paralysis on small decisions while ensuring rigor for strategic decisions.
bun run ${CLAUDE_SKILL_DIR}/scripts/select_depth.ts --budget 100000 --impact medium --reversibility moderate
# Output: STANDARD tier (1-2 days, full 4 phases)
Phase 1: Classify (Core vs Context)
Purpose: Determine if capability is Core (competitive differentiator) or Context (table stakes).
CORE CONTEXT
Competitive differentiator Table stakes
Unique to business model Industry-standard
Source of customer value Required but not differentiating
Customers choose you BECAUSE of this Customers assume you have this
Amazon's recommendation engine Auth systems
Netflix's content algorithm Payment processing
Uber's matching system Email delivery
Default: BUILD Default: BUY
(unless constraints prevent) (unless no viable vendors)
Integration: Run cynefin-classifier first if problem domain unclear. Complex domains favor Buy/Partner to reduce risk.
Outputs:
- Core or Context classification with 3+ supporting reasons
- Strategic importance score (1-10)
- Red line criteria (Never Build / Never Buy boundaries; see detailed Red Line Criteria section below)
Exit Criteria:
- [ ] Classification documented with justification
- [ ] Stakeholder consensus (or disagreement documented)
- [ ] Strategic importance score assigned
See: references/core-vs-context.md for detailed framework
Phase 2: Analyze (TCO + Capacity)
Purpose: Quantify total cost of ownership and assess team capacity.
bun run ${CLAUDE_SKILL_DIR}/scripts/calculate_tco.ts \
--build-initial 500000 \
--build-ongoing 100000 \
--buy-initial 50000 \
--buy-ongoing 200000 \
--partner-initial 100000 \
--partner-ongoing 150000 \
--discount-rate 0.12 \
--years 5
| Category | Build | Buy | Partner |
|---|---|---|---|
| Initial | Engineering time, design, architecture, tooling | License fees, implementation services, integration, training | Integration work, legal agreements, POC |
| Ongoing | Maintenance, bug fixes, features, infrastructure, support | Subscription fees, support contracts, upgrades | Revenue share, relationship management, co-dev coordination |
| Hidden | Opportunity cost, onboarding, tech debt | Vendor lock-in, integration debt, workflow constraints | Roadmap misalignment, partner dependency, revenue complexity |
Capacity Assessment Questions:
- Does team have required skills? (If no, add training/hiring costs)
- Does team have capacity without delaying strategic projects?
- Can team maintain long-term? (Avoid orphaned code)
- Does build option develop strategic capabilities for future?
Exit Criteria:
- [ ] TCO calculated for 3, 5, 10 year horizons
- [ ] Sensitivity analysis identifies top 3 cost drivers
- [ ] Team capacity assessment documented
See: references/tco-methodology.md for calculation details
Phase 3: Evaluate (Decision Matrix)
Purpose: Score options across strategic, operational, and risk dimensions.
bun run ${CLAUDE_SKILL_DIR}/scripts/score_decision.ts \
--criteria-file "decision-criteria.json" \
--options "build,buy,partner"
# Output:
# Strategic: BUILD=7.2 BUY=5.8 PARTNER=6.1
# Operational: BUILD=5.5 BUY=7.8 PARTNER=6.0
# Risk: BUILD=6.0 BUY=6.5 PARTNER=5.5
# Winner: BUILD (confidence: MEDIUM, 9% gap)
| Dimension | Weight | Criteria |
|---|---|---|
| Strategic | 40% | Alignment, Market signaling, Optionality, Asymmetric upside |
| Operational | 30% | Time to value, Team fit, Integration complexity, Maintenance burden |
| Risk | 30% | Vendor risk, Execution risk, Regulatory risk, Lock-in risk |
Integration: Run pre-mortem on leading option. If >5 severe risks surface, reconsider decision.
Exit Criteria:
- [ ] All criteria scored for all options (with justifications)
- [ ] Pre-mortem completed for leading option
- [ ] If tie (<10% gap), document tie-breaker rationale
See: templates/decision-matrix.md for scoring worksheet
Phase 4: Decide (Final Decision + ADR)
Purpose: Make final decision, document rationale, plan reassessment.
| Option | When to Choose |
|---|---|
| Build | Core capability + team capacity + favorable TCO. No viable vendors OR vendor lock-in unacceptable. Strategic capability development desired. |
| Buy | Context capability + viable vendors + faster time to value. Team lacks capacity OR skills. Commodity capability with mature market. |
| Partner | Shared value creation (rev share, co-development). Strategic alliance benefits beyond technology. Neither build nor buy individually compelling. |
| Defer | Unclear requirements OR high uncertainty. Market immature (wait for consolidation). Problem may not need solving (validate demand first). |
Integration:
- Feed final rationale to decision-critic for validation
- Trigger adr-review for multi-agent consensus (Standard/Deep tier)
Exit Criteria:
- [ ] Decision documented in ADR (Standard/Deep tier)
- [ ] Decision-critic validation passed (no blocking issues)
- [ ] Reassessment schedule established (6, 12, 24 months)
- [ ] Stakeholder sign-off obtained
Scripts
calculate_tco.ts
Calculate NPV, IRR, break-even timeline for build/buy/partner options.
bun run ${CLAUDE_SKILL_DIR}/scripts/calculate_tco.ts --help
# Required flags:
# --build-initial FLOAT Initial build costs
# --build-ongoing FLOAT Annual build costs
# --buy-initial FLOAT Initial buy costs
# --buy-ongoing FLOAT Annual buy costs
# --partner-initial FLOAT Initial partner costs
# --partner-ongoing FLOAT Annual partner costs
# --discount-rate FLOAT Discount rate (0.10 = 10%)
# --years INT Analysis horizon (3, 5, or 10)
# Exit codes:
# 0: Success
# 10: Warning (negative NPV detected)
# 11: Error (missing cost categories)
score_decision.ts
Calculate weighted decision scores with sensitivity analysis.
bun run ${CLAUDE_SKILL_DIR}/scripts/score_decision.ts --help
# Required flags:
# --criteria-file PATH JSON file with criteria weights and scores
# Exit codes:
# 0: Clear winner (>20% score gap)
# 1: Tie requires human judgment (scores within 10%)
check_reassessment_triggers.ts
Detect assumption drift, recommend re-evaluation.
bun run ${CLAUDE_SKILL_DIR}/scripts/check_reassessment_triggers.ts --help
# Required flags:
# --adr-file PATH ADR markdown file
# --current-state PATH JSON with current assumptions
# Exit codes:
# 0: Assumptions hold, stay course
# 10: Minor drift (<20%), monitor closely
# 11: Major drift (>20%), re-evaluation required
score_vendor.ts
Score vendor stability, pricing, feature fit.
bun run ${CLAUDE_SKILL_DIR}/scripts/score_vendor.ts --help
# Required flags:
# --vendor-data PATH JSON file with vendor information
# Exit codes:
# 0: Pass (score >70)
# 10: Yellow flag (score 50-70)
# 11: Red flag (score <50)
Templates
All templates available in templates/ directory:
core-vs-context-analysis.md- Phase 1 classification worksheettco-analysis.md- Phase 2 TCO calculation worksheetdecision-matrix.md- Phase 3 multi-criteria scoringadr-buy-vs-build.md- Phase 4 ADR template
Anti-Patterns
| Avoid | Why | Instead |
|---|---|---|
| Using framework for trivial decisions (<$10K, reversible) | Analysis paralysis | Use judgment, skip framework |
| Skipping pre-mortem on high-risk decisions | Misses hidden failure modes | Always run pre-mortem in Phase 3 |
| Not documenting rationale | Future team can't understand why | Create ADR (Standard/Deep tier) |
| Forgetting reassessment plan | Decision becomes stale | Set calendar reminders for reviews |
| Optimizing for single dimension (only cost) | Ignores strategic value | Use full decision matrix |
Verification
- [ ] All four phases completed with documented outputs
- [ ] TCO analysis covers 3-year horizon minimum
- [ ] Decision matrix scored by 2+ stakeholders
- [ ] Risk assessment includes top 3 risks per option
- [ ] Final recommendation includes reassessment triggers
Related Skills
| Skill | Relationship |
|---|---|
| cynefin-classifier | Pre-step to determine problem domain |
| pre-mortem | Phase 3 risk identification |
| decision-critic | Phase 4 validation |
| adr-review | Phase 4 multi-agent consensus |
| planner | Post-decision execution planning |
<details> <summary><strong>Deep Dive: ADR Template</strong></summary>
# ADR-NNN: [Build/Buy/Partner/Defer] {Capability Name}
## Decision
We will [BUILD/BUY/PARTNER/DEFER] {capability}.
## Context
{Problem being solved, strategic importance, business drivers}
## Considered Options
1. **Build**: {Summary, pros/cons, TCO}
2. **Buy**: {Summary, pros/cons, TCO, vendor options}
3. **Partner**: {Summary, pros/cons, TCO, partnership models}
4. **Defer**: {Summary, validation approach}
## Decision Drivers
1. {Top factor that determined outcome}
2. {Second most important factor}
3. {Third most important factor}
## Consequences
### Expected Outcomes:
- {Positive outcome 1}
- {Positive outcome 2}
### Risks:
- {Risk 1 + Mitigation}
- {Risk 2 + Mitigation}
## Reassessment Triggers
- Cost assumption changes >20%
- Time horizon shifts materially
- Strategic priority shifts (context to core or vice versa)
- Vendor viability concerns (M&A, financials, EOL)
- Team capacity changes (key departures, hiring surge)
- Competitive dynamics shift (urgency increases)
- Regulatory changes
- Technology disruption
- Customer demand signal
- **Annual review:** {Date}
## Decision-Maker
{Name, Title} on {Date}
</details>
<details> <summary><strong>Deep Dive: Reassessment Plan</strong></summary>
# Run quarterly to detect assumption drift
bun run ${CLAUDE_SKILL_DIR}/scripts/check_reassessment_triggers.ts \
--adr-file "architecture/ADR-123-build-payments.md" \
--current-state "current-state.json"
# Output:
# Status: REASSESSMENT_REQUIRED
# Drift: Vendor pricing changed 25% (trigger: >20%)
# Recommendation: Run Phase 2 (TCO) analysis with updated costs
</details>
<details> <summary><strong>Deep Dive: Red Line Criteria</strong></summary>
Never Build:
- Commodity, undifferentiated, no strategic value
- Team lacks skills
- Time-critical delivery
Never Buy:
- Core IP, competitive secret sauce
- 100% control required
- No viable vendors
- Regulatory restrictions
</details>
<details> <summary><strong>Deep Dive: Timelessness (9/10)</strong></summary>
Why timeless:
- Economic fundamentals of make-vs-buy unchanged for 100+ years (industrial revolution to present)
- Strategic concepts (core vs context) from Wardley Mapping remain relevant 20+ years later
- Decision structures universal across software, hardware, services
- Human cognitive biases (sunk cost, confirmation bias) do not change
- Principal/VP decision-making authority stable across decades
What could change (-1 point):
- AI code generation could shift build costs radically downward
- Vendor consolidation (only 1-2 vendors) changes buy calculus
- Open source maturity creates "free" middle ground
Mitigation: Framework includes "Defer" option and reassessment triggers to adapt to these shifts.
</details>
References
Deep-dive documentation in references/ directory:
core-vs-context.md- Core vs Context framework from Wardley Mappingtco-methodology.md- TCO calculation best practicespartnership-models.md- Partnership structures (co-dev, revenue share, alliance)vendor-evaluation.md- Vendor assessment frameworkreassessment-playbook.md- How to re-evaluate when assumptions change
04 同仓库更多 Skills
validation-authority
Treat upstream validators as authoritative. Align local config to them. Use when validation fails unexpectedly, before modifying validator behavior, or when tempted to change upstream tool code.
using-brain-memory
Guidance for using Brain semantic memory effectively. Applies Zettelkasten atomic note principles with observations and relations. Use when deciding whether to query or create notes, structuring note content, or understanding the semantic knowledge graph.
threat-modeling
Structured security analysis using OWASP Four-Question Framework and STRIDE methodology. Generates threat matrices with risk ratings, mitigations, and prioritization. Use for attack surface analysis, security architecture review, or when asking what can go wrong.
taste-lints
Custom lints with agent-readable remediation instructions. Enforces taste invariants (file size, naming conventions, structured logging, complexity) and surfaces errors that agents can act on directly. Use when writing or reviewing code to catch style violations early.
style-enforcement
Validate code against style rules from .editorconfig, StyleCop.json, and Directory.Build.props. Detects line ending violations, naming convention issues, indentation problems, and charset mismatches across C#, Python, TypeScript, and JavaScript. Produces JSON reports for pre-commit hooks and CI pipelines.
steering-matcher
Match file paths against steering file glob patterns to determine applicable steering guidance. Use when orchestrator needs to inject context-aware guidance based on files being modified.