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 / Do NOT use

Use this skill: "should we build or buy a customer data platform, partner with a vendor, or defer the investment for two years", "strategic sourcing decision for our identity provider with TCO". Every one of these compares multiple sourcing options over a multi-year horizon with budget stakes.

Do NOT use this skill for: "is there a library for parsing YAML in Python" (narrow, tactical, one implementation choice; use programming-advisor instead). The dividing line is time horizon and option count: a multi-option strategic trade-off with TCO stays here, one tactical answer goes to programming-advisor.

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 plan instead if:

• Sourcing decision already made, need milestones and tasks to build it

Quick Reference

Process

Phase 0: Depth Selection

Select appropriate analysis depth to prevent over-engineering.

Why: Time-box effort proportional to decision magnitude. Prevents analysis paralysis on small decisions while ensuring rigor for strategic decisions.

Example: A $100K budget with medium impact and moderate reversibility matches the Standard row on all three axes, so the depth is Standard: 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:

1. Core or Context classification with 3+ supporting reasons
2. Strategic importance score (1-10)
3. 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

Use: the classification prompts in this phase for the Core vs Context framework.

Phase 2: Analyze (TCO + Capacity)

Purpose: Quantify total cost of ownership and assess team capacity.

python3 scripts/calculate_tco.py \
  --build-initial 500000 \
  --build-ongoing 100000 \
  --buy-initial 50000 \
  --buy-ongoing 200000 \
  --partner-initial 100000 \
  --partner-ongoing 150000 \
  --discount-rate 0.12 \
  --years 5

Capacity Assessment Questions:

1. Does team have required skills? (If no, add training/hiring costs)
2. Does team have capacity without delaying strategic projects?
3. Can team maintain long-term? (Avoid orphaned code)
4. 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

Use: the TCO categories and sensitivity checks in this phase for calculation details.

Phase 3: Evaluate (Decision Matrix)

Purpose: Score options across strategic, operational, and risk dimensions.

python3 scripts/score_decision.py \
  --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)

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

Use: the scoring table in this phase as the decision worksheet.

Phase 4: Decide (Final Decision + ADR)

Purpose: Make final decision, document rationale, plan reassessment.

Integration:

1. Feed final rationale to decision-critic for validation
2. 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.py

Calculate NPV, IRR, break-even timeline for build/buy/partner options.

python3 scripts/calculate_tco.py --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.py

Calculate weighted decision scores with sensitivity analysis.

python3 scripts/score_decision.py --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.py

Detect assumption drift, recommend re-evaluation.

python3 scripts/check_reassessment_triggers.py --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.py

Score vendor stability, pricing, feature fit.

python3 scripts/score_vendor.py --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

This skill ships no separate template files. Each phase worksheet lives in this document:

• Phase 1 classification worksheet: the Core vs Context grid and exit criteria in "Phase 1: Classify (Core vs Context)"
• Phase 2 TCO worksheet: the cost category table in "Phase 2: Analyze (TCO + Capacity)", computed by calculate_tco.py (see Scripts)
• Phase 3 decision matrix: the weighted dimension table in "Phase 3: Evaluate (Decision Matrix)", scored by score_decision.py (see Scripts)
• Phase 4 ADR template: the "Deep Dive: ADR Template" section below

Anti-Patterns

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

# 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}

# Run quarterly to detect assumption drift
python3 scripts/check_reassessment_triggers.py \
  --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

References

Deep-dive documentation in references/ directory:

• references/PHASE1_ANALYSIS.md - Core vs Context classification worksheet
• references/SKILL_SPEC.xml - Original skill specification. Its template and reference inventories list files that were never shipped.