Product Investment Analysis

Priority: P0 (CRITICAL)

Evaluate feature and product investment options with rigorous incremental accounting, cash versus capacity boundaries, and deterministic financial calculations without fabricating speculative returns.

1. Capability Preflight

Before numeric evaluation:

• Confirm Node >=20, the installed package's scripts/evaluate-investment.mjs, and assets/investment.schema.json separately. Invoke with node <installed-root>/scripts/evaluate-investment.mjs <absolute-input.json>; resolve companion assets from the package, never the monorepo root or current directory. Do not install missing resources.
• Resolve ledger capability from <agent-skill-root>/product-management/product-evidence/SKILL.md (ordinary install), <agent-skill-root>/product-management-product-evidence/SKILL.md (flattened Kiro install), or MCP get_skill("product-management", "product-evidence"). Distinguish explicit exclusion from missing pack; if configuration cannot establish status, say unknown.
• Missing script/schema/runtime is a resource/runtime gap; missing accounting data means unknown or qualitative-only analysis; missing budget authority permits proposal only, never commitment. If ledger capability is absent, continue standalone only when supplied inputs do not require ledger validation.
• Never simulate ledger validation, infer financial completeness, or call an analysis approved without required human authority.

2. Core Protocol

1. Strategic Alternatives Evaluation:
   • Explicitly evaluate five investment paths: build, buy, partner, defer, and do-nothing.
   • Model the "do nothing / defer" counterfactual as the baseline against which incremental costs and benefits are measured.
2. Trade-offs & Opportunity Cost:
   • Assess displaced work and opportunity costs qualitatively in the decision rationale.
   • Do NOT record opportunity costs as cash expenses unless approved by finance.
3. Cash vs Capacity Distinction:
   • Treat developer and operational hours saved as operational capacity (basis: "capacity_hours").
   • Capacity hours are informational. Never convert saved hours into cash benefits ($B$) unless an explicit finance-verified cash reduction mechanism exists.
4. Finance-Owned Incremental Accounting Boundary:
   • Restrict analysis to incremental cash flows within the declared accounting_boundary.
   • Exclude genuinely sunk costs and non-incremental historical overhead allocations.
   • Finance determines and documents whether specific step-fixed costs or incremental shared contracts are included relative to do-nothing. The calculator cannot infer cost completeness or causality.
5. Scenario Rigor (Base, Downside, Upside):
   • Formulate separate, independent assumptions for base, downside, and upside scenarios.
   • Do NOT average scenarios or assign arbitrary probabilities to compute an expected return.
   • An invalid scenario rejects the entire analysis; do not drop downside risks silently.
6. Qualitative Investment Validity:
   • Qualitative investments (compliance, platform stability, strategic enablement) are valid without fabricated numerical ROI.
   • If benefits are non-monetary, declare nonmonetary: string[] and leave monetary benefits unknown.
7. Budget-Holder Authority & Revisit Triggers:
   • Economic evaluation models potential value; it does NOT authorize capital expenditure or team allocation.
   • All investments require formal budget-holder authority binding before commitment.
   • Define unambiguous revisit triggers (cost overruns, missed milestones, macro shifts).

3. Anti-Patterns & Red Flags

• No Fabricated ROI: Never invent dollar figures for unquantifiable strategic or user-satisfaction benefits.
• No Capacity as Cash: Never treat saved internal labor hours as immediate cash revenue or cash savings.
• No Sunk Cost Inclusion: Never justify continuing an investment based on money already spent.
• No Probability-Weighted Blending: Never average downside and upside to mask capital loss scenarios.
• No Double-Counting: Never net COGS in contribution margin and simultaneously claim them as explicit costs.

4. References

• Financial Contract — Exact arithmetic, rounding, payback, and NPV specification.
• Examples & Reference Oracles — Baseline, zero-cost, durable payback, and rejection fixtures.
• Manual Semantic Rubric — Case-specific financial decision/artifact criteria; lexical assertions are not semantic verdicts.
• Citations and Sources — Authorship statement, conceptual citations, and exclusions.