Standards
Research standards
The requirements every dossier meets before it is delivered, how its confidence rating is computed, and the defined limits of scope.
Delivery gate
Four checks, applied to every dossier
Each dossier is checked against the following before release. A failure on any one holds delivery until it is resolved.
| Check | Standard | Definition |
|---|---|---|
| Citation resolution | 100% | Every citation must resolve to a retrievable document. A single unresolved citation holds delivery. |
| Sourced findings | 90% minimum | Findings carry a numbered citation to the document behind them. Anything below threshold is reviewed by an analyst before release. |
| Confidence calibration | No overstatement | The confidence a dossier claims is checked against the authority of the evidence it cites. |
| Primary record coverage | Analyst verified | Where a target files with the SEC or appears in federal court records, those records are present in the evidence base. |
Calibration
How the confidence rating is computed
Every dossier carries a confidence rating derived from the authority of the sources cited behind its findings. Filings and court records weigh more heavily than secondary coverage. The computed rating is reconciled against the confidence the dossier claims for itself before delivery.
High
Findings rest principally on filings, court records and the company's own disclosures.
Medium
A primary record exists but is partial — the standard outcome for a private company.
Low
Little of substance is on the public record. The dossier states this and scopes its findings accordingly.
Unresolved questions reduce the rating independently of source quality. A dossier on a listed company may therefore be rated medium. Ratings are computed from the evidence base and are not adjusted by hand.
Validation
How the pipeline is tested
The research pipeline is benchmarked against a fixed control set of eight targets with independently established profiles. Three are listed companies filing audited accounts. Three are private, where the evidentiary ceiling is lower and a dossier claiming otherwise is scored as a failure.
Two are null controls. No public record exists for either, so a substantive finding against them registers as an error the benchmark is designed to catch. Runs that fall below threshold on citation resolution, source authority or calibration are not released.
Scope
Defined limits
- Coverage is delivered on a defined cadence. A retainer comprises four briefs per month; continuous monitoring between briefs is not included.
- Private targets are covered from the records that exist. Most lower-middle-market companies file little or nothing, and the confidence rating reflects the available evidence base.
- Expert networks, licensed data rooms and confidential information are outside scope. Output is analysis, and is not investment, legal or tax advice.
- Engagements are scoped before they are accepted. Where a target cannot be identified on the public record, the engagement is declined at intake.
Corrections
Correction policy
Where a delivered finding is not supported by its cited source, the dossier is corrected and reissued at no charge, and the client is notified directly. Corrections are logged below with the date and the change made.
Correction log
No corrections issued. Log opened 19 August 2026.
Verification
Tracing a finding to its source
Every finding is tied to a numbered source, and the index at the back of the dossier identifies the document behind it. Filings carry their SEC accession number, court records their docket number, and archived pages their capture date. Each opens on the issuing body’s own website.
Where sources conflict, the dossier presents both readings and states which is better evidenced.