Everything-PR's Earned Media Originality Study found that two of 52 classified coverage URLs for 10 corporate announcements contained identifiable original reporting, and both belonged to one story family. The audit located 57 eligible URLs across nine announcements. It is an exploratory pilot, so the figures describe this sample only and do not estimate how much journalism is original.
Last updated: November 8, 2026. Research completed November 8, 2026. Study type: exploratory structured pilot.
Summary of findings
Everything-PR examined whether coverage of a press release adds anything beyond the release itself. Of 52 classified URLs, 24 were republications, eight were rewrites, 16 added context, two contained original reporting and two stayed uncertain. Executive quotations repeated in 39 of 52 URLs, and no article-specific interview was identifiable in any of them.
Table of contents
A coverage report can contain many URLs and very little new reporting. It can also contain a syndicated article that preserves real reporting done by its originating publication. Counting placements alone cannot tell those outcomes apart.
What did the audit count?
The audit counted coverage of 10 preselected corporate announcements across six sectors: technology, consumer products and beauty, food and beverage, healthcare, financial services, and travel and hospitality. Each announcement received the same six-query search procedure, and coverage was eligible from the announcement date through 14 days afterward, inclusive.
The sample has two announcements each in technology, consumer and beauty, food and beverage, and healthcare, plus one each in financial services and hospitality. All were published September 1 to 9, 2026, inside a wider September 1 to October 15 eligibility period.
The original project sought 60 announcements. After substantial checkpoint research, its scope was reduced with authorization. The completed study keeps the entire preselected pilot rather than choosing ten announcements with attractive outcomes. The other 50 candidates are outside this study, and they are not observations with zero coverage.
| Classification | URLs | Share of 52 classified |
|---|
| Republication | 24 | 46.2% |
| Rewrite | 8 | 15.4% |
| Added context | 16 | 30.8% |
| Original reporting | 2 | 3.8% |
| Uncertain | 2 | 3.8% |
The searches located 57 eligible coverage URLs across nine announcements. Five URLs could not receive a whole-article classification because their bodies were unavailable or incomplete, which leaves the 52 shown above. At announcement level, nine of 10 had eligible coverage, eight had at least one URL with added context or original reporting, and one had identifiable original reporting.
For SugarbearPRO's Hair Prenatal announcement, no eligible coverage was located. Product listings, corporate pages, a mirrored original and an unrelated guide to another vitamin product did not qualify. That result describes the search procedure and does not establish that no coverage exists anywhere.
Can original reporting survive syndication?
Original reporting can survive syndication, and the Ghirardelli case shows it. Parade's September 1 article describes an editor seeing the chocolates in a grocery store and supplies retailer-specific availability details. That is firsthand verification absent from the announcement, and a new interview is not required when an article documents another form of reporting.
The AOL version carries the same reporting and explicitly credits Parade. The audit therefore records two URLs, one story family and one originating reporting contribution. It does not credit AOL with a second grocery-store check.
That makes 2 of 52 classified URLs, or 3.8%, and 1 of 37 families with assessable contribution, or 2.7%. Both denominators include the unresolved classifications. Neither percentage estimates the share of corporate coverage that is original.
The Parade article also repeats a marketing executive's statement from the release. A carried-over quote and new reporting can sit in the same piece, so the classification follows the material addition and the feature ledger records the repeated quote separately.
Disclosure applies to this case. Ghirardelli's original release names 5W Public Relations as media contact. Everything-PR is published by Everything-PR News LLC, which is owned by a shared interest group whose ownership overlaps with 5W's. The case received the same rules as every other announcement, its distribution copies remained republications, and the retail observation alone justified the original-reporting classification. The study makes no attribution of that reporting to agency performance.
How much coverage adds context?
Sixteen of the 52 classified URLs added context without identifiable original reporting. Together with the two original-reporting URLs, 18 of 52 classified URLs, or 34.6%, materially extended the announcement under the study's rules. The additions included prior company developments, product background and operational analysis.
Reinsurance News' Corgi Re coverage links the launch to earlier company developments. That helps a reader place the announcement in sequence, and it does not establish that the publication obtained a new interview for the article.
Several foodservice articles added advice about seasonal menus, catering and merchandising. Those analytical additions qualify as context even without independently identifiable reporting. The audit does not endorse every commercial assertion in that analysis, because a new-to-baseline fact is an assertion absent from the announcement and not a certification that it is correct.
These categories describe observable contribution. They do not measure audience value, fairness, commercial impact or a ranking of journalism, and republication can accurately communicate an announcement.
Do outside sources mean the publication did its own reporting?
An outside source does not prove independent reporting. Fourteen of 52 classified URLs, or 26.9%, contained an observed source outside the announcing company, yet none of the 52 contained an interview identifiable as conducted for that article.
The two numbers coexist because an outside voice can arrive through the release itself. A doctor quoted in a medical-device release remains a person outside the announcing company when a publication repeats the quotation, and the quotation does not become a new interview under an editorial headline.
In the ElectroWire case, Cardiac Rhythm News reproduced supplied clinical details and physician and founder remarks. It was classified as a rewrite, and its outside-source feature was recorded separately. The audit found no evidence of an article-specific interview in that body.
The absence of identifiable interview evidence is narrower than a claim that nobody spoke to a reporter. The study classifies the public evidence it inspected and cannot reconstruct undocumented newsroom processes.
How often do executive quotations repeat?
Executive quotations repeated in 39 of 52 classified URLs, or 75.0%. The table shows how the repetition spread across contribution categories.
| Classification | URLs with repeated quote | URLs classified |
|---|
| Republication | 21 | 24 |
| Rewrite | 5 | 8 |
| Added context | 10 | 16 |
| Uncertain | 1 | 2 |
| Original reporting | 2 | 2 |
Quote presence is therefore a weak originality test. A quote can survive a full release reproduction, an extensively reworded brief or a piece that adds verification. The useful question is where the quotation originated and what the article added around it.
A different quotation is also insufficient to establish a new interview. In one Corner Bakery item on CourierPR, executive wording differed from the baseline and another quotation appeared without identifiable interview provenance. The study kept an uncertain classification rather than awarding original reporting or alleging fabrication.
A Bake Magazine Corner Bakery item likewise stayed uncertain because it introduced a different availability date without naming a source. Uncertainty is an appropriate finding when the evidence cannot separate additional reporting from an unsupported alteration or another source version.
How did publication dates change the results?
Publication dates decided which coverage qualified. The Shiseido serum review located on Marie Claire was published September 24 and fell outside that announcement's window, so it was excluded. A Broadway News article about the Algonquin renovation was also too late, while several hotel stories appeared before the selected September 8 baseline.
Those earlier hotel articles expose a boundary of the design: a selected press release may not be the first public communication about an event. The audit compares later coverage with that baseline and does not trace the first disclosure of any event.
Five classified articles carried modification dates after their coverage windows. They qualified on their initial publication dates, but the study inspected current public versions and not archived snapshots. Removing those five leaves 47 classified URLs with the same two original-reporting URLs, or 4.3%. That sensitivity check does not prove when every surviving sentence was first published.
Does an AI label say anything about originality?
An AI label describes disclosure and says nothing about originality by itself. One eligible Kazia coverage page on StockTitan explicitly labels AI-generated analysis and questions and answers alongside a reproduced release. The added explanatory material supports an added-context classification, and the AI disclosure is recorded as a separate field.
The classification does not validate those explanations as medical or investment guidance. It also does not treat a stock-market scanner as proof that a journalist independently verified the clinical announcement. Generic profile widgets and unrelated recommendation modules were excluded from contribution coding.
The audit draws no inference about AI authorship from writing style, a publisher's name or repetitive phrasing. "Not observed" means no explicit disclosure was found in the inspected material, not that AI assistance was absent.
What can communications teams do with these findings?
Communications teams can change what a coverage report records. A useful ledger keeps distribution status, full-text access, story family and evidence of the article's contribution, and it matches each executive quote against the announcement with a short rationale for every apparent addition.
That structure prevents four common counting errors:
- Treating a wire copy as independent coverage.
- Counting an original article and its syndication as two reporting events.
- Reading a supplied expert quote as a fresh interview.
- Scoring an inaccessible article as a rewrite because its opening resembles the release.
The findings do not justify telling clients that a particular percentage of journalism is unoriginal. They support a narrower reporting change: show what was distributed, what was rewritten, what was contextualized and what was independently reported, and leave unresolved evidence visible. Teams planning a distribution program can compare options in Everything-PR's press release distribution services guide.
How was the study run, and what are its limits?
The study ran three fixed queries per announcement through each of two search engines and kept up to the first ten hits per query. The 60 query records contained 504 ranked hits and 235 announcement and URL candidate records before further consolidation. Social posts, indexes, product pages, unrelated items and original-release mirrors were excluded.
Coverage was eligible from announcement day through day 14 inclusive, and a cap of 12 eligible article URLs per announcement applied; none reached it. Every classification used inspected source text. Of 57 eligible URLs, 52 were classified and five stayed unclassified for access or completeness: a Klaus Way Acquis article with no readable body, a Nation's Restaurant News Corner Bakery slideshow with missing slides, an FNBX Ghirardelli preview, a Nasdaq Kazia article with truncation and an AM Best Corgi article behind a paywall.
Five other article candidates had unverified publication days and sit outside the eligible-window denominators: Dealroom and Owler for Acquis, The Vault for Shiseido, Allrecipes for Ghirardelli and Newsmaker Alert for the Algonquin. They are not zero-originality observations.
All uncertain classifications and 10 of the other 50 were reopened for a same-researcher consistency check. The 12 checks produced four feature-field corrections across three articles and no originality-class changes. No independent reviewer or inter-rater reliability is claimed.
The sample is small, was selected through a changing search-visible frame and is concentrated in early September. The pilot also helped settle the coding rules, so the results are exploratory and not a confirmatory test on fresh cases. Publisher versions can change, search results can omit articles, related sites can share ownership and access failures may conceal reporting. Sector cells hold one or two cases each, so the study offers no sector rankings and makes no claims about complete pickup, audience reach, causation, ROI or business impact.
For earlier work on the same theme, see Everything-PR's press release pickup study, whose three cases fall outside this sample's announcement period. More studies sit in the Everything-PR Research Index and the research section.
