A materiality assessment in filings, in the forensic sense we apply, quantifies whether a company's narrative claims in its public filings are supported by audit-traceable public-record evidence. A successful assessment yields a claim, an evidence map, a numeric reconciliation, and a verdict qualified by explicit thresholds, not a subjective impression. We treat this discipline as the backbone of evidence-first corporate accountability work, including the methodology we apply at Lacuna Index.
TL;DR:
- Record each claim’s wording, source URI, retrieval time, evidence map, numeric reconciliation, materiality bands, assumptions, and revision trail so another analyst can reproduce the verdict.
- Use SEC structured APIs for XBRL facts instead of scraping rendered filings; log query parameters and exhibit file hashes to preserve a reproducible audit trail.
- Check adjusted figures for recurring restructuring costs or stock compensation, and compare press releases with later filings for changed figures and unexplained reclassifications.
- A documented mismatch cannot establish intent or reveal undisclosed liabilities, and enforcement may arrive years later, so monitor new filings and SEC comment letter exchanges.
Table of Contents
- Which public sources establish the evidentiary record?
- How do you run a reproducible assessment, claim to verdict?
- What red flags tend to precede a restatement or enforcement action?
- What tools make this kind of check scalable and auditable?
- What are the limits of this kind of verification?
- A practitioner's note on operationalizing this workflow
- How Lacuna Index turns this workflow into a subscription product
- FAQ
- Sources
Which public sources establish the evidentiary record?
Forensic reconciliation begins with EDGAR, the SEC's foundational public record of corporate disclosure, because every claim that matters eventually touches a form that the SEC collects there. We prioritize sources by evidentiary weight rather than by how prominently a company features them in its own communications.
- CORRESP filings — preserve the back-and-forth between SEC staff and company counsel, showing how a disclosure was challenged or revised before it reached its final form.
Programmatic access through data.sec.gov's RESTful APIs and daily or bulk index files makes this evidentiary record auditable at scale. Rather than scraping rendered HTML, we query structured endpoints that return the same submission history and XBRL-tagged facts a regulator would see, which matters because reproducibility is the entire point of forensic work rather than an incidental benefit.
How do you run a reproducible assessment, claim to verdict?
A materiality review process only earns the word "forensic" when another analyst can rerun it and reach the same conclusion. We structure the work into five steps, each with a defined input, output, and audit record.
- Capture the claim. Record the exact quote, its source URI, and a retrieval timestamp, then classify the metric type: narrative assertion, percentage, non-GAAP figure, or forward-looking statement. A vague claim ("disciplined cost management") is classified differently than a specific one ("adjusted EPS grew 14%").
- Map the required evidence. Identify which XBRL facts, footnotes, exhibits, or CORRESP references could confirm or contradict the claim. This step produces an evidence map, a short list of exactly where a disinterested reader would need to look.
- Reconcile the numbers. Where a company claims a non-GAAP EPS uplift, pull the corresponding GAAP EPS figure from XBRL and the disclosed adjustments line by line. If the adjustments quietly exclude recurring restructuring charges, the reconciliation will show it.
- Score the gap. Compare the reconciled figure against documented materiality bands rather than intuition, and record every assumption made in scoring. This is the step that converts a discrepancy into a graded verdict instead of an accusation.
- Monitor and update. Subsequent filings and SEC comment letters can change the picture, so the verdict carries a revision date and an audit trail of every URI and query used to build it.
Pro Tip: Store a file hash alongside every exhibit you retrieve: filings occasionally get amended or superseded, and the hash lets you prove which version your reconciliation was built on.
This sequence is the practical answer to how to assess materiality without relying on reputation or press coverage as a proxy for performance. Readers building their own version of this process can find a fuller walkthrough of the claim-to-GAAP reconciliation step in our forensic analysis workflow.
What red flags tend to precede a restatement or enforcement action?
Certain patterns recur often enough across SEC enforcement history that they function as early filing materiality criteria, warning signs worth checking before a position is sized or a governance recommendation is written.
- Non-GAAP adjustments that habitually exclude recurring costs, such as routine restructuring or stock-based compensation reclassified as "one-time," distort the picture a reader takes away from the headline figure.
- Repeated reclassifications or dependence on large, infrequent cash adjustments buried in footnotes rather than addressed in the MD&A narrative.
- Aggregated percentage claims with no stated methodology, such as an ESG integration figure cited without a written policy or scope definition.
- Inconsistent metrics across document types, where the press release attached to an 8-K uses a different figure than the 10-Q filed weeks later.
- CORRESP exchanges showing staff pushback on a specific disclosure, which often precedes a quiet revision in the next filing cycle.
SEC staff guidance on non-GAAP measures identifies non-GAAP adjustments as the most frequent source of misleading narrative claims, and enforcement actions involving companies including DXC, Bausch Health (formerly Valeant), FCA US, and Invesco each trace back to a variant of these patterns: an adjustment, a reclassification, or an aggregated claim that did not survive reconciliation against the underlying facts. None of these cases prove intent on their own, but each one shows why the flags above are worth checking before, not after, a figure becomes a headline.
What tools make this kind of check scalable and auditable?
Manual scraping does not scale, and it leaves no clean provenance trail. The EDGAR application programming interfaces expose structured endpoints, including company-concept and companyfacts, that return XBRL-tagged figures directly, removing the parsing ambiguity that comes from pulling numbers out of rendered PDF tables.
- Query companyfacts for a single issuer to retrieve every XBRL fact the company has tagged across filing history in one structured call.
- Cross-reference CORRESP and comment-letter histories to see which disclosures attracted staff questions and how the company's language shifted afterward.
- Use bulk ZIP archives when building sector-level comparisons rather than issuing thousands of individual requests.
Pro Tip: Log the exact query parameters used for every API call alongside the returned data, not just the result; a reconciliation that cannot be rerun with the same inputs is not an audit trail, it is a snapshot.
What are the limits of this kind of verification?
Public-record forensic analysis can show divergence between a claim and the evidence; it cannot prove internal intent or surface liabilities that were never disclosed anywhere. GAO's analysis of disclosure practices notes that determining whether limited disclosure reflects "nothing material happened" or "insufficient disclosure" is often impossible from the public record alone, which is why findings should be written in probabilistic language rather than as accusations. Enforcement is also a lagging indicator: cases often surface years after the underlying conduct, so proactive monitoring of filings and CORRESP exchanges does more to reduce exposure than waiting for a regulatory outcome.
A practitioner's note on operationalizing this workflow
We built a system to run this exact sequence at scale: every claim in a filing becomes a row in an evidence log, tied to an XBRL fact, an exhibit, or a CORRESP reference, and scored into what we call an execution score. Users can leverage forensic reports, sector benchmarks, and evidence maps to support governance decisions and published analysis, each one traceable back to a specific filing and timestamp. Reproducibility is the standard we hold ourselves to, not a marketing claim.
— Glen
How Lacuna Index turns this workflow into a subscription product
We sell the output of this exact process, not a black box. Per-company analysis applies the claim-to-verdict workflow above to a single issuer, producing a forensic report with an execution score and a documented evidence map; see the benchmarks page for coverage. Cohort Pulse and Sector Sweep, detailed on our pricing page, extend the same methodology across peer groups and full sectors, including Financial Services and Industrials, for readers who need cross-issuer comparability rather than a single report.
- Per-company analysis: a single-issuer forensic report with an audit-traceable evidence map.
- Cohort Pulse: narrative-versus-delivery comparison across a defined peer group.
- Sector Sweep: benchmark coverage across an entire sector, built on the same scoring logic.
Full methodology, including our scoring bands and evidence standards, is documented on our methodology walkthrough. Readers preparing filings for a transaction may also find value in a partner resource on data room documentation standards, which addresses disclosure readiness from the preparer's side. Start by reviewing sector coverage, then request access through Lacuna Index to see a full report.
FAQ
What does a materiality assessment in filings actually measure?
It measures whether the specific claims a company makes in its public filings are supported by audit-traceable evidence, such as XBRL facts, exhibits, or CORRESP records, rather than whether the company is "good" or "bad" in a general sense. The output is a reconciled gap between narrative and delivery, not a sentiment score.
How is this different from a standard financial statement analysis?
Standard analysis typically evaluates financial health using ratios and trends; this forensic review process instead traces specific narrative claims, including forward-looking or non-GAAP statements, back to their supporting public-record evidence. The focus is claim-level, not statement-level.
What is the single most common red flag in filing materiality criteria?
Non-GAAP adjustments that exclude recurring costs are the most frequently cited pattern in SEC enforcement history, because they let a company present an adjusted figure that diverges meaningfully from the audited GAAP result.
Can this kind of analysis prove a company misled investors?
No. A forensic assessment can document a divergence between a claim and the available public evidence, but proving intent or uncovering undisclosed liabilities requires non-public records that this methodology does not have access to, as GAO's disclosure analysis makes clear. Findings should be stated as probabilistic, evidence-qualified conclusions.
Where do I find the raw data to run this kind of check myself?
Start with EDGAR's data access documentation for the RESTful APIs and bulk indexes, then cross-reference the SEC's CORRESP submission guide for comment-letter context. Both are free and form the backbone of the workflow described above.