review-and-validate

Review and Validate

When to use

For strategy office, FP&A, or scorecard owners running periodic BSC reviews. Scoring measures alone misses the point — this skill tests whether causal links from build-strategy-map held in the data, separating execution failure from theory failure.

What this skill does not do

• Does not build or revise the strategy map — recommends build-strategy-map revisit when theory problems are diagnosed; does not rewrite mechanisms inline.
• Does not set targets or select measures — reads existing scorecard artifacts only.
• Does not invent actuals — missing data → INPUT NEEDED / structured first pass.
• Does not replace OKR retros — BSC causal-link testing is distinct from quarterly KR scoring.

Preconditions

Input	If missing
Live strategy map (profile location)	Halt — produce structured first pass with INPUT NEEDED
Current-period measure actuals vs targets	List gaps; do not invent values or causal verdicts
Practice profile scoring convention	Use default RAG with [PROVISIONAL] tag

Provisional mode

When map or period data is incomplete:

• Label CONFIDENCE: structured first pass.
• Score only measures with data; mark links NOT TESTABLE THIS PERIOD where leading data absent.
• Do not assign CONTRADICTED without evidence.

Trust spine

• Confidence bands (governance-tracking):
  • High: Data current and instrumented; every measure scored; every link tested with lag rules applied.
  • Medium: Partially current data or [PROVISIONAL] scoring convention; some links NOT YET OBSERVABLE.
  • Low: Stale or missing actuals — flag explicitly; no causal verdicts presented as fact.
• Tag vocabulary: [verify], [review], [PROVISIONAL], [unverified — planning convention] on default lag windows.
• Failure modes:
  • Strategic advice vs. support: Diagnoses execution vs theory with evidence; recommends map revisit — strategist decides whether to refresh.
  • Client confidentiality: Review may contain proprietary performance data — CONFIDENTIAL header; destination check.
  • Accountability gap: CONTRADICTED links require execution-vs-theory split with evidence — not a single RAG rollup that hides theory failure.
  • Analytical Rigor: N/A — governance review, not MECE decomposition.
  • Incentive Gaming: Guards against RAG-washing (green leading measures while lagging outcomes flat) and execution-default bias (blaming teams when theory is contradicted). Requires explicit theory-problem call when conditions met; early map refresh trigger prevents deferring uncomfortable map revisions.
• Escalation triggers:
  • Two+ theory-problem links or one on load-bearing top link → recommend early map refresh.
  • Measure definition changed mid-period → flag execution diagnosis unreliable.
  • Mixed evidence on CONTRADICTED link → inconclusive; name data needed next period.
  • External confound on leading measure → execution problem or NOT TESTABLE; do not confirm theory.