What this note is
This note is the companion to the firm’s full validation reference at /methodology/validation/. The validation page enumerates the twenty-two canonical VERIDEX gates and the four engine extensions. This note does the second-order work: it explains what to do with the gates as a reader.
A principal receiving a VERIDEX-validated deliverable is not asked to audit the work. The audit has already happened. The principal is asked to read the work knowing where the gates apply and what they protect against. This note is the field manual for that reading.
The six integrity families
The twenty-two canonical gates organise into six integrity families. Each family does a specific job.
1. Sourcing — five gates
G·01 figures trace to source · G·02 institutions verifiable · G·03 statistics named · G·19 source tier disclosed · G·20 temporal provenance tagged
The Sourcing family asks: Where did the numbers come from, and when? It exists to defend against the most common analytical failure — figures introduced without provenance, used as if they were ground truth.
Where to look in a report: any sentence containing a numeric claim. Each such sentence should carry a citation that resolves to a primary source, identifies the source’s tier, and tags the date the source was published or updated.
2. Analytical — five gates
G·04 no advocacy · G·05 different question = different content · G·06 score-narrative alignment · G·17 absorption scored against historical · G·18 political-economy analytical not partisan
The Analytical family asks: Is the work answering the question that was asked, or a different question? It defends against scope creep and the slow drift toward advocacy in long documents.
Where to look: the relationship between the development question on the first page and the conclusion on the last. The conclusion should answer the question, not a related question.
3. Modelling — four gates
G·08 scenario parameters consistent · G·11 Monte Carlo correlation explicit · G·12 three-step decomposition (TFS→AFS→RDS / TAM→SAM→SOM) · G·15 arithmetic verifiable
The Modelling family asks: Does the math work, and would it work the same way if a different analyst ran it? It defends against the most subtle class of error — internally inconsistent parameter sets and undisclosed correlation assumptions.
Where to look: anywhere a probability distribution, a market sizing or a multi-period projection appears. Each such artefact should disclose the parameters and let the reader recompute.
4. Risk — four gates
G·07 risk profiles specific · G·09 Counterfactual Pause · G·10 Opportunity Cost ≥ 0.5% / $50M · G·16 counterfactual ≥ 2 named alternatives
The Risk family asks: What is the cost of being wrong, and what was considered and rejected? It defends against the most common pathology of recommendation work — the analyst’s bias toward the preferred path, unaccompanied by an honest accounting of what the path costs to abandon.
Where to look: any recommendation. Each recommendation should carry an explicit counterfactual, an opportunity-cost figure, and at least two named alternatives that were rejected.
5. Consistency — two gates
G·13 Decision-Maker CTA discipline · G·14 Impact ↔ Risk consistent
The Consistency family asks: Does the report contradict itself? It defends against documents whose risk section says one thing and whose recommendation section says another.
Where to look: cross-references between sections. Risk profiles and recommendations should be quantitatively reconcilable.
6. Presentation — two gates
G·21 Executive Summary 6 components equal weight · G·22 Decision-Maker Briefing separated
The Presentation family asks: Is the artefact shaped so the principal can use it? It defends against the most banal failure — a report that is right but unusable.
Where to look: the Executive Summary and the Decision-Maker Briefing. Gate 21 mandates the Executive Summary’s six components carry equal weight. Gate 22 mandates the Decision-Maker Briefing live in its own artefact, separate from the analysis.
The engine extensions — four gates
G·24 Rate-currency · G·25 Source-quality + temporal-provenance · G·26 Addressable-market decomposition · G·27 Executive Summary construction
The four engine extensions are gate-level constraints specific to the firm’s three analytical engines. They are not part of the canonical twenty-two but they are part of the firm’s working standard. The engine they apply to is named in each deliverable’s audit trail.
A note on Gate 23. Gate 23 is reserved for a future extension. The firm has not invented it. The deliberate gap is part of the standard.
What the gates do not do
The gates do not produce certainty. They produce a clear audit trail of where certainty stops and judgement starts. The principal still has to make the call.
The firm’s position is that this is what an honest deliverable looks like — that the alternative, in which the gates are skipped or hidden, produces the appearance of certainty without the substance. VERIDEX exists to make the substance visible.
Cross-references
- Full gate enumeration. /methodology/validation/
- The framework that produces the work the gates validate. /methodology/framework/
- The decisions where the gates were applied. /case-studies/