Research Standards

Every numerical claim is reconstructed from at least three independent evidence paths drawn from five canonical evidence classes. Classes are weighted in this order when paths disagree:

• Regulatory filings, audited corporate disclosures filed under regulatory obligation. The strongest evidence class available.
• Official statistics, sovereign and multilateral statistical authorities. Used as macroeconomic anchors for sanity-checking market trajectories.
• Industry benchmarks, association and trade-body datasets that capture operator-level economics inside a defined sector perimeter, weighted by membership coverage.
• Syndicated benchmarks, a curated set of independent published market estimates used for directional triangulation. Where our figure diverges, the divergence is disclosed, never averaged away.
• Primary research, analyst-conducted conversations with operators, buyers, and regulators. The only class that captures qualitative context behind the numbers.

The full five-stage framework that turns these classes into a published figure is documented on the methodology page.

No figure publishes without committee sign-off by a human reviewer. The committee is adversarial by design and independent of the authoring analyst. Each review produces a written red-team memo that ships alongside the report, not edited out. Roles on the committee:

• Editorial Committee, final publication authority; signs every report.
• Sector editor, desk-level review for industry frame, figure plausibility, and landscape accuracy.
• Red-team reviewer, deliberately outside the authoring team; writes the red-team memo.
• Methodology reviewer, audits evidence paths, triangulation integrity, and compliance with the framework.
• Corrections steward, owns the post-publication corrections ledger and the revision audit trail.

Every report is produced under the research responsibility of a named human analyst. Proprietary tooling supports synthesis, cross-checking, and coherence-testing stages under analyst supervision. Every figure is written, reviewed, and defended by a human. The tooling does not have autonomous publication authority.

If a reader identifies a suspected error in a published figure, email corrections@meridianconsensus.com with the page URL, the specific claim, and the evidence supporting the correction. The corrections desk acknowledges within one business day and resolves within three.

Confirmed corrections are handled as follows:

• Material corrections, any change to a published figure, ranking, or directional thesis: fixed in place, with a dated correction note appended to the report, the prior figure preserved, and the dateModified field updated. An entry is added to the public errata log on the About page.
• Non-material corrections, typos, formatting, broken links: fixed silently, but the correction log still records the edit.
• Disputed claims, where our figures diverge from independently published benchmarks, the divergence is disclosed inline, never quietly reconciled to the median.
• Retractions, if a published figure is found to be materially wrong and cannot be repaired, the figure is retracted with a dated notice on the page. Retractions are never removed from the errata log.

• Preview pages (/markets/{slug}), refreshed every 90 days by default.
• Event-triggered revisions, a top-10 company 10-K filing, major M&A in the sector, regulatory action, or material capital-markets event triggers an earlier refresh. The triggering event is logged in the revision note.
• Commissioned reports, refreshed on commission, on client request, or when an event-trigger affects the underlying thesis. Clients receive a written diff memo against the prior edition on every refresh.

Meridian Consensus does not accept paid placements. Company rankings, segment shares, and competitive positioning are produced entirely from public data and our validation pipeline. We do not accept sponsorship, editorial payments, or ranking improvements in exchange for money or consideration of any kind.

Analysts disclose personal financial interest in any company profiled. Where disclosure is incompatible with maintaining market integrity, the analyst is recused from that study.

Commissioned-research clients do not have editorial control over the figures, thesis, or conclusions of a report. Clients see the committee-signed output; they do not negotiate it.

Readers may quote excerpts of up to 300 words per report in presentations, memos, and filings with clear attribution to Meridian Consensus and the page URL. Longer excerpts or syndication require written permission. Redistribution of the full report without a multi-seat licence is not permitted, see the Terms of Service.

Buyers of commissioned reports have unlimited written Q&A access with the named author during the licence period. All correspondence is archived and searchable across the client’s seats. Readers of free preview pages can flag factual concerns through the corrections address above.

We follow a preview-first policy. Every market preview, the methodology, and a complete specimen report are free to read end-to-end before commissioning. Commissioned reports are final on delivery. If a delivered report is materially inaccurate or does not match the disclosed scope on the preview page, email research@meridianconsensus.com, the analyst team reviews case-by-case (partial credit, re-run, or refund where warranted). See the Refund Policy for the full terms.

This policy applies to every page on this site, every commissioned report, every advisory memo, and every export delivered under the Meridian Consensus name. Updates to the policy are recorded on the About page errata log and announced in the research digest.