§16.8 · Part X transparency · v16.8.0
Measurement methodology
The §16.8 doctrine commits AIVAON to publishing the methodology behind every externally-reported category-leadership metric. External stakeholders — investors, board members, analyst community, press, and prospective customers — must be able to inspect the measurement substrate independently to validate the rigor that supports our category-leadership claims. Below is the full methodology stack: 13 binding principles, index normalization grammar, redaction rules, Measurement Council governance, and mutation workflow.
Thirteen binding measurement principles
Principle 1
Truth precedence over comfort
Measurement is designed to produce truthful signal even when uncomfortable; missing data does not earn credit.
Principle 2
Definitional precision
Every metric carries a rigorous definitional specification covering what is measured, how, what inclusions/exclusions apply, what aggregation produces the reported value, and what temporal windows govern measurement.
Principle 3
Competitive comparability
Metrics are defined to permit competitive comparison where competitive data is available.
Principle 4
Leading-indicator integration
The scoreboard includes leading indicators alongside lagging indicators so feedback is timely enough for corrective action.
Principle 5
Cohort-based measurement
Metrics are measured across customer cohorts to distinguish performance on customers acquired at different times and under different conditions.
Principle 6
Segmentation depth
Metrics are decomposable across the segments that matter commercially — market cell, customer tier, vertical, use case, acquisition channel.
Principle 7
Frequency calibration
Measurement frequency matches the metric's signal latency and the decision timeline the metric informs.
Principle 8
Trend integration
Metrics are presented with temporal trend visualization rather than as point-in-time snapshots.
Principle 9
Target specification
Every headline metric carries explicit targets with the target basis documented.
Principle 10
Anomaly detection and investigation
Metric anomalies trigger structured investigation that produces root-cause understanding.
Principle 11
Cross-metric correlation analysis
The measurement infrastructure supports correlation analysis across metrics that produces causal understanding.
Principle 12
Attribution rigor
Metric improvements are attributed to the specific investments and actions that produced them.
Principle 13
Governance-grade rigor
The measurement infrastructure satisfies the rigor standards that would withstand external audit.
Index normalization grammar
The Category Leadership Index is a composite 0–100 indicator produced from 26 normalized metric sub-scores. Each metric is normalized direction-aware against its current-quarter target:
- higher_is_better → normalized = clamp(value / target × 100, 0, 100)
- lower_is_better → normalized = clamp(target / value × 100, 0, 100); value ≤ 0 yields full credit (machine zero).
- target_band → normalized = 100 × (1 - distance_from_band_center / band_width), clamped to [0, 100].
- Missing data → 0 sub-score. Truth precedence over comfort: gaps do not earn credit.
Per-dimension sub-scores are weighted by the per-dimension weights (commercial 40%, strategic 30%, product 20%, technical 10%); the composite is the weighted sum across dimensions. Full weight set is captured in every snapshot for reproducibility.
Redaction rules for external report packs
Each external report pack (investor / board / analyst / press / prospect) declares its own set of redaction rules to strip competitively-sensitive content before distribution. The redaction engine is declarative + auditable; every pack generation persists the list of redactions applied.
| Rule | Behaviour |
|---|---|
| redact_per_cell_breakdowns | Removes per-market-cell breakdown values from each metric in the pack payload. |
| redact_named_competitor_ids | Strips competitor reference IDs from per-metric annotations. |
| redact_revenue_absolute_values | Replaces USD-denominated metric values with a redacted-marker so percentages are still useful but absolute revenue is not disclosed. |
| redact_pricing | Drops pricing-related metrics (PRICING + ACV) from the pack entirely. |
| redact_pipeline_values | Replaces pipeline-related metric values with a redacted-marker. |
| redact_competitor_displacement_rate | Drops competitor displacement metrics from the pack entirely. |
| redact_competitor_clusters | Removes the competitor cluster summary from the pack envelope. |
| redact_per_competitor_displacement | Removes per-competitor displacement breakdown from the pack envelope. |
Measurement Council governance
Every change to a metric definition, leadership index weight, cohort definition, report pack, or source-reconciliation policy flows through the Measurement Council. Proposals walk the state machine: proposed → under_review → approved_shadow → activated (or retired, rejected, withdrawn).
- Quorum: 5 of 7 council members must vote.
- Approval: more approve votes than reject votes.
- Shadow window: 30 days between approval and activation; the new definition runs alongside the old so the historical-comparability impact can be observed before cutover.
- Audit trail: every proposal + vote + transition recorded permanently with full attribution.
The full Measurement Council declaration is at /api/v1/aivaon/scoreboard/council.
Review cadences
The §16.8 measurement loop runs on four cadences each linked into a §16.7 governance cadence so the scoreboard feeds the Execution Operating System rather than running parallel to it.
weekly
Weekly Scoreboard Review
Reviews: dim_commercial, dim_product, dim_technical
monthly
Monthly Scoreboard Review
Reviews: dim_commercial, dim_product, dim_technical, dim_strategic
quarterly
Quarterly Scoreboard Ratification
Reviews: dim_commercial, dim_product, dim_technical, dim_strategic
annual
Annual Scoreboard Trajectory Review
Reviews: dim_commercial, dim_product, dim_technical, dim_strategic
Cadences past their interval × 1.25 grace window auto-emit alerts so the absence of a cadence execution becomes a signal, not a silent gap.
Mutation workflow
- Doctrine mutations land via PR to
backend/artifacts/aivaon-section16.8-category-scoreboard.json. - The §16.8 CI gate (10 jobs) enforces JSON Schema validity, structural invariants, service + runtime smoke, MCP catalog uniqueness, migration chain integrity, backend↔frontend mirror sync, routes import smoke, extension invariants, extension migration sanity, and extension service smoke.
- The runtime never mutates the artifact; operator interactions write only to the §16.8 ledger tables.
- Every reading + cadence + annotation + alert resolution + index snapshot + benchmark + cohort assignment + proposal + vote + pack generation + approval + distribution carries the recording principal for audit attribution.
- Reporting corrections issue a new generation pointing back at the original via
correction_of; the original is markedcorrectedfor audit.