Press release analysis, in the forensic sense, is the discipline of treating a voluntary corporate announcement as a set of testable claims rather than as news. Each claim is inventoried, reconciled against the required filing it accompanies, evaluated for emphasis and metric definition, and then tracked against later public disclosures to establish whether delivery followed. The approach used at Lacuna Index applies this process consistently across issuers, using public records only, so that the resulting evidence is traceable and comparable.
TL;DR
- A press release is voluntary and unaudited in tone. The Form 8-K it often accompanies is the required document, and the two frequently differ in language, so reconciliation is the first test.
- Headline performance numbers and positive tone move prices initially, then reverse: academic evidence shows abnormal returns followed by drift, which is why emphasis should be scored rather than trusted.
- Red flags cluster around undefined metrics, unsupported superlatives, missing comparison periods and promises with no later reporting line.
- A red flag is a prompt for further testing, not proof of misconduct. Readability, incremental information and ordinary promotional habit all produce false positives.
- Delivery is established only by tracing each claim forward through filings, calls and proxy statements.
Table of Contents
- What press release analysis means for a forensic reader
- Why the press release is not the filing
- A five-step method
- Press release red flags worth scoring
- Accountability metrics for release claims
- Limits: when a red flag is not evidence
- How Lacuna Index applies the method
- Perspective
- FAQ
- Sources
What press release analysis means for a forensic reader
Most pages that rank for this phrase address marketers: reach, impressions, headline construction and distribution. That is the sender's view. The reader of a forensic evaluation occupies the opposite position, asking what the issuer wants the market to believe, what the filed record supports, and what remains unstated.
Three questions structure the analysis:
- What was claimed? Quantities, comparisons, timelines, rankings and commitments, each isolated as a discrete statement.
- What does the required record show? The corresponding Form 8-K, 10-Q or 10-K, and the transcript of any related earnings call.
- What happened next? Whether later disclosures confirmed, restated, quietly dropped or reframed the claim.
The third question is the one most often skipped, and it is the one that converts a reading exercise into an accountability measure. An analysis that stops at tone describes the narrative. An analysis that follows the claim forward measures the aspiration-to-execution gap.
Why the press release is not the filing
A company that experiences a material event frequently files a Form 8-K, the current report required by the SEC, and may also issue a press release. A study of U.S. public firms found that 37% of the time firms voluntarily issue a press release about an event already covered by a required 8-K, and that they are more likely to do so when the 8-K tone is positive. The same study attributes part of the tonal difference between the two documents to quotes from firm officers.
This matters for two reasons. First, the selection of which events receive a second, more favorable telling is itself information. Second, readers appear to treat the two documents as interchangeable. Using EDGAR search-traffic data from 2008 to 2011, the authors detected a 30 to 50% decrease in unique 8-K searchers when a press release accompanies the 8-K, and an 80% decrease for 8-Ks released simultaneously with a press release. Fewer people read the filed document precisely when a more promotional version is available.
The price effects are measurable. The study reports abnormal returns of 88 to 109 basis points around the event window for each standard-deviation increase in press release sentiment bias, then a negative drift of roughly 13 to 20 basis points over the following month. That pattern is consistent with investors initially responding to both information and spin. A basis point is one hundredth of a percentage point.
For a step-by-step treatment of the filing side of this comparison, see the guide to 8-K filing analysis for U.S. analysts.

A five-step method
Step 1: Build a claim inventory
Read the release once without evaluating anything and extract every statement that could later be verified. Classify each as one of four types:
- Quantitative result (revenue, margin, cash flow, customer counts).
- Comparative claim (growth versus a prior period, versus peers, versus guidance).
- Forward commitment (targets, launch dates, savings programs).
- Qualitative superlative (leading, record, best-in-class, unprecedented).
The inventory should record the exact wording, the paragraph, and whether an executive quote or the issuer's own narrative voice makes the claim. Quotes deserve separate tagging because the evidence above links them to tonal divergence from the filed text.
Step 2: Reconcile against the required document
Place the release beside the 8-K exhibit, or the periodic report when the release summarizes results. For each inventoried claim, record whether the filed document (a) states the same fact, (b) states it with different qualifiers, (c) omits it or (d) contradicts it. Differences in category (b) and (c) are the most common and the most informative: a figure that appears in the release without its definition or comparison base in the filing has been moved from a controlled context to an uncontrolled one.
A framework proposed in an ISB working paper on voluntary disclosure quality offers a useful structure here. It assesses decision-usefulness, applies quality attributes from the literature, and derives red-flag metrics measurable in real time at the firm-quarter level, covering non-GAAP earnings, management guidance and conference calls. The paper's approach supports the principle that release analysis should produce a repeatable list of measurable flags rather than an impression.
The timing of release relative to filing is part of this step. Patterns in who learns what and when are covered in the guide to detecting selective disclosure patterns within 24 hours.
Step 3: Evaluate emphasis, headline numbers and superlatives
Emphasis is quantifiable. A study of more than 17,000 U.S. earnings releases over eleven years, summarized by the American Accounting Association, found that about 28% specified performance numbers in headlines. In the sample, 72.1% of releases had no performance statistic in the headline, 11.9% had one, 8.5% had two, 2.9% had three and 4.5% had more than three.

The consequences are instructive. Adding one strong performance number increased the three-day results-driven price boost by an extra one-third, followed by a 60-day reversal greater than the initial boost, and firms that headlined strong results showed lower earnings persistence. Persistence was measured by whether year-to-year gains were equaled or exceeded in a third year.
A practical scoring routine for superlatives:
| Test | Question | Evidence that passes |
|---|---|---|
| Referent | Compared with what? | A named peer set, period or prior guidance figure |
| Metric | Defined where? | A definition in the filing, with reconciliation if non-GAAP |
| Base period | Is the comparison period stated? | Explicit year-ago or sequential base |
| Support | Does the filed record contain the underlying data? | Matching figure in the 8-K, 10-Q or 10-K |
| Repeatability | Was the same claim made in prior quarters? | Consistent wording and metric across periods |
A superlative that fails the referent and support tests is scored as unsupported emphasis. One that passes both is scored as supported, however promotional the vocabulary. Readability should be recorded separately: an Accounting Review experiment found that when guidance and year-ago benchmarks are inconsistent, higher readability of positive trend performance leads to more favorable investor judgments. Clear writing can therefore amplify an inconsistently benchmarked claim.
Step 4: Test metric definitions
The most consequential release failures involve a metric that means less than it appears to. In November 2020, the SEC charged former Wells Fargo executives John Stumpf and Carrie Tolstedt over allegedly misleading investors about the bank's cross-sell metric, which was inflated by unused, unneeded or unauthorized accounts and services. The allegations illustrate why a headline ratio should be traced to its construction: what counts in the numerator, and whether the denominator is stable.
For adjusted earnings measures, the same discipline applies. The 6-point checklist for scoring non-GAAP adjustment quality sets out how to evaluate recurring "one-time" items and the transparency of the bridge to GAAP.
Step 5: Track delivery forward
Every forward commitment from Step 1 enters a ledger with a date. At each subsequent 10-Q, 10-K, earnings call and proxy statement, the ledger is updated with one of four statuses: delivered, delayed with explanation, silently dropped, or redefined. The share of commitments in each status across several quarters is the core accountability metric; it is also the measure that an individual release can never provide. Related techniques for comparing narrative with the annual filing are described in the guide to annual report discrepancy in equity research, and the broader logic in how public disclosures reveal management gaps.
Press release red flags worth scoring
The following patterns recur in enforcement records and in the academic literature. Each is a scoring input rather than a conclusion.
- Undefined or novel metrics. A proprietary indicator introduced without a definition or history.
- Track-record claims without verification. In a 2015 action, the SEC found that Virtus Investment Advisers advertised a substantially overstated performance track record dating to April 2001, when the strategy's sub-adviser admitted no client assets had tracked it from April 2001 to September 2008. The affected funds' assets grew from $191 million at the end of 2009 to about $11.5 billion by 2013.
- Repetition across channels. In its complaint against Solucorp Industries, the SEC alleged the company claimed in at least ten press releases, filings, an annual report and a shareholder letter to hold contracts generating over $350 million in contingencies. Identical claims repeated across documents are not corroboration; they are one claim.
- Segment results without disclosed adjustments. The SEC's 2014 action against CVS alleged overstated results for a retail pharmacy line and failure to disclose accounting adjustments tied to an acquisition.
- Tone that outruns the filing. A release materially more positive than its 8-K, with the difference carried by executive quotes.
- Headline stacking. Multiple performance numbers in a headline, which the AAA study associates with lower persistence.
- Reputation repair after a restatement. Research by Ed deHaan found that announcements of executive replacements, governance revamps and control improvements after intentional misreporting lifted share prices on average 2%, and that 51% of 898 post-restatement announcements targeted nonfinancial constituencies. Such announcements merit the same claim-to-delivery tracking as any other.
Accountability metrics for release claims
A consistent metric set allows releases to be compared across issuers and quarters. The following five are the most portable:
- Support rate: the share of inventoried claims with a matching figure in the required filing.
- Definition coverage: the share of metrics with a stated definition and, where relevant, a GAAP reconciliation.
- Tonal gap: the difference in sentiment between the release and its 8-K.
- Delivery rate: the share of dated commitments reported as delivered by their stated horizon.
- Persistence: whether stated gains are sustained in the following period, following the AAA study's definition.
The market does price credibility failures. A PLOS ONE study of 1,269 newspaper articles on reported misconduct at 462 U.S.-listed firms between January 1995 and January 2020 found cumulative abnormal returns of -4.1% when misconduct is reported, with larger penalties when the corporation rather than an individual was blamed and when article tone was negative. Scoring delivery before such an event is the purpose of an accountability metric set. Cross-issuer interpretation requires a peer reference, which is why benchmarking disclosure quality across sectors is the natural extension of single-release work.
Limits: when a red flag is not evidence
Release analysis produces hypotheses. Several findings counsel restraint.
First, a release may carry real information. The Kellogg study found that investors respond to both information and spin, and that press releases may contain incremental information beyond the 8-K. A tonal gap is therefore a reason to read more closely, not a verdict.
Second, difficult text is not necessarily deceptive text. A study of annual earnings announcements by U.S.-listed firms from 2007 to 2018 found that lower readability produces higher unexplained trading volume and abnormal volatility, a form of investor disagreement, while financial analysts help reduce that disagreement and low readability was unassociated with post-announcement returns. Opacity belongs in a disclosure-quality score, not automatically in a misconduct inference.
Third, enforcement cases are calibration, not base rates. The SEC actions cited above are allegations or findings in specific matters, and most releases with a superlative or a favorable adjustment involve nothing of the kind.
Fourth, the method depends on a comparator. A listed issuer has an 8-K, a proxy statement and a transcript to reconcile against. For companies outside the public-reporting regime, no required document exists, and the analyst must first locate reliable private company financial data before any claim can be tested, which is a materially weaker position than the public-company case.
How Lacuna Index applies the method
Lacuna Index measures the gap between corporate narrative claims and actual delivery using public records only: earnings calls, SEC filings, press releases, proxy statements and public statements, processed with one methodology across every covered company. Sector benchmarks are free to read, starting with Financials, and the methodology walkthrough documents the scoring. Per-company forensic reports, with execution and narrative scores and audit-traceable evidence, are available by subscription through the per-company analysis offering, and the Cohort Pulse and Sector Sweep plans are described on the pricing page.

For an analyst or journalist who would otherwise assemble the claim inventory, filing reconciliation and delivery ledger manually for each issuer, the standardized process supplies a consistent score and a governance log for each result.
Perspective
The recurring error in release analysis is to read the document as prose when it functions as an instrument. A release is built to be read quickly, and the evidence shows that it is: filed documents attract fewer readers when a more favorable telling is available alongside them.
Our position is that the unit of analysis should be the claim, not the release, and that a claim is only resolved when a later public record confirms or fails it. That is why we measure the aspiration-to-execution gap across quarters, with the same rules for every issuer, and why every score carries a governance log that a reader can audit.
Public records are sufficient for this work. No insider access is needed to see that a metric was introduced without a definition, that a commitment disappeared between two calls, or that a superlative had no referent. What is required is consistency, and a method that does not change depending on the company being examined.
Glen
FAQ
What is the first thing to check in a press release?
Extract each verifiable claim, then compare it with the Form 8-K or periodic filing covering the same event. The filed document is the required record, so any figure, qualifier or comparison base that differs from it is the first place to look.
Do superlatives in a press release count as red flags?
Only when they lack a referent or filed support. Research on more than 17,000 earnings releases links strong performance numbers in headlines to a short-lived price boost, a larger 60-day reversal and lower earnings persistence, so emphasis is worth scoring, but a superlative backed by a defined metric and a matching filing is simply supported.
Can a press release red flag prove misconduct?
No. A red flag indicates that further testing is warranted. Releases can contain incremental information, and poor readability is associated with investor disagreement rather than with post-announcement returns, so conclusions require reconciliation with filings and later disclosures.
Which accountability metrics are most useful?
Support rate, definition coverage, the tonal gap between release and 8-K, delivery rate on dated commitments, and persistence of stated gains. Together they convert a single release into a measurable record that can be compared across quarters and issuers.
Sources
- Information or Spin? Evidence from Language (Kellogg School of Management)
- American Accounting Association: Take heed when companies flaunt (October 8, 2018)
- The Accounting Review: consistency of benchmark performance
- ISB working paper: Beyond Earnings Quality
- SEC: charges against former Wells Fargo executives (2020)
- SEC: Virtus advertised false performance (2015)
- SEC: charges against CVS (2014)
- SEC complaint: Solucorp Industries
- Words and numbers: a disagreement story from post-earnings announcements
- Stanford GSB: Ed deHaan on rebuilding trust after a scandal
- PLOS ONE: Reported corporate misconducts and financial markets
