What Sources Do Wikipedia Company Articles Actually Cite?

Why ask this question

Wikipedia's notability guideline tells you what sources are supposed to establish a company's place in the encyclopaedia: significant coverage, in reliable sources, independent of the subject. What it does not tell you is what the articles that already exist are actually built from.

That gap matters commercially, because the guideline is what a buyer is told and the citation list is what an editor sees. If the two diverge, advice based only on the guideline will misprice which coverage is worth pursuing.

So we took 24 existing company articles and extracted every citation in them. The sample is deliberately ordinary: consumer and technology companies of the kind that buy media placement, including Stripe, Canva, Figma, Databricks, Duolingo, Rivian, Chobani, Patagonia, Warby Parker, Instacart, DoorDash and Peloton.

One methodological point shapes everything that follows. A large share of Wikipedia citations point at the Internet Archive rather than the publisher, so a naive count returns web.archive.org as the top source. We unwrapped every archived URL to credit the original publisher, which is the only way the numbers mean anything.

What the citations actually are

Across 24 articles we extracted 2,465 citation URLs spanning 548 distinct domains. The concentration at the top is striking: TechCrunch alone accounts for 244 citations, roughly one in ten of everything cited.

Behind it the list is recognisably the business press: CNBC 110, The Wall Street Journal 78, The Verge 69, The New York Times 68, Forbes 65, Bloomberg 53, Reuters 48, Fast Company 46, Business Insider 45 and Fortune 34.

Primary and regulatory sources appear in meaningful numbers too. SEC filings account for 24 citations, and company domains appear throughout — Rivian 33, the Duolingo blog 26, Databricks 19, Stripe 10, Canva 8, Chobani 8, Figma 7.

The long tail is where most of the diversity sits. With 548 domains supporting 2,465 citations, the median domain is cited once or twice: trade titles, regional outlets, specialist blogs like Electrek and InsideEVs, and one public radio station. Company articles are not built from a shortlist of prestige mastheads.

How those sources are rated

We then cross-referenced the cited domains against our parse of Wikipedia's own perennial sources list. Of the 2,465 citations, 952 have a domain that maps to a rated entry — 38.6% of the total, with the remainder going to outlets the list has never had cause to adjudicate.

Within that rated group the split is closer than the guideline's language would suggest. Generally reliable sources take 497 citations, or 52.2%. Sources rated no consensus take 403, or 42.3%. Generally unreliable sources take 52, or 5.5%.

In other words, among citations to sources Wikipedia has an opinion about, nearly as many go to sources it has not endorsed as to sources it has. That is not a scandal — no consensus means contested, not forbidden, and such a citation can be perfectly appropriate for an uncontroversial fact — but it is a long way from the tidy picture of articles resting on unimpeachable sourcing.

The single largest contributor to that 42.3% is TechCrunch, whose entry warns that its articles “often have too little independent analysis” to satisfy the depth requirement for corporate coverage, and advises weighing how far a piece relies on the subject's public relations material.

There is a limit to what a domain-level analysis can see, and it cuts against the comfortable reading rather than for it. Wikipedia rates Forbes in three separate entries, splitting staff reporting from contributor articles, but a citation's domain is forbes.com either way. The same is true of HuffPost and Rolling Stone. So our 65 Forbes citations contain an unknown mixture of material Wikipedia considers generally reliable and material it considers self-published and generally unreliable, and we have counted all of them under the publication's headline no-consensus rating.

The self-citation share

Nine per cent of all citations — 223 of 2,465 — point at the company's own website, blog or newsroom. That is a larger share than most people expect given how firmly the notability guideline excludes non-independent sources.

There is no contradiction. Independence is required to establish that a subject merits an article at all; it is not required for every sentence inside one. A company's own site is an acceptable source for uncontroversial descriptive facts such as founding dates, product names, headquarters or executive titles.

The distinction is worth internalising because it is routinely misused in sales conversations. Your own material can support the furniture of an article. It cannot support the case that the article should exist, and no volume of it substitutes for independent coverage.

The same logic explains the 24 SEC citations. Regulatory filings are primary sources, excellent for verifiable financial facts and useless for demonstrating that anyone independent found the company significant.

What is missing from the list

Two absences are as informative as the rankings. The first is press releases. Across 2,465 citations, PR Newswire appears 11 times and no other wire appears in any meaningful volume — and PR Newswire is rated generally unreliable, on the stated ground that its releases are “not subject to editorial oversight”.

For a category of content produced in enormous quantity and sold specifically on its distribution, near-total absence from the citation base of company articles is the clearest available evidence of what it is worth for this purpose.

The second absence is any entry at all for several of the most-used outlets. CNBC at 110 citations, Fast Company at 46 and Fortune at 34 have no perennial sources entry, meaning editors have never needed a standing argument about them. Being absent from that list is closer to a quiet endorsement than a gap: contentious sources are what get listed.

That also caps what our rating analysis can claim. It describes the 38.6% of citations whose domain we could map to an entry, and says nothing about the majority that sit outside it.

It is also worth sitting with the size of the unrated majority. Of 2,465 citations, 1,513 — 61.4% — point at domains with no perennial sources entry at all: regional business journals, specialist trade titles, sector blogs, a public radio station, YouTube at 18 citations. These are the ordinary working sources of company coverage, and Wikipedia has simply never needed a standing position on them. Any analysis of this kind describes the contested minority and leaves the bulk of real sourcing unexamined.

What this means for anyone buying coverage

The most useful finding for a buyer is about outlet selection. Technology and business trade press — TechCrunch, The Verge, VentureBeat, CNBC, Fast Company — carries a far larger share of the citation base for company articles than the prestige national titles that dominate tier-one pricing.

But volume is not standing. The most-cited domain in our sample is rated no consensus, and its entry warns specifically about reliance on public relations material. Coverage that reads as processed announcement can be cited for a fact while doing nothing to establish that a company matters.

If the objective is a Wikipedia article, the realistic target is independent journalists at outlets rated generally reliable writing something substantive that did not originate with you. Everything else in this dataset — self-citations, filings, wire releases, contested trade coverage — furnishes an article that already exists rather than bringing one into being.

Media Morpheus sells media and podcast placement, including wire distribution. That wires appear 11 times in 2,465 citations is an unflattering number for part of what we sell, and it is reported here because it is what the data shows.

Methodology and limitations

On 24 September 2026 we retrieved the raw wikitext of 24 English Wikipedia company articles via the MediaWiki raw action and extracted every external URL. URLs wrapped by the Internet Archive or archive.today were unwrapped so that the original publisher is credited rather than the archive; without this step web.archive.org is the single largest host in the data. We excluded Wikimedia-internal links and DOI resolvers. Domains were normalised by stripping the www prefix. Cross-referencing to source ratings used a manually constructed map from the most-cited domains to entries in our 18 September 2026 parse of the perennial sources list; 952 of 2,465 citations mapped, and all rating percentages describe that mapped subset only. Self-citation was counted by matching each article's subject to its own domain. Limitations: 24 articles is a purposive sample of consumer and technology companies that buy media placement, not a random sample of Wikipedia; citation counts measure use within existing articles and not what caused those articles to be accepted; and a domain-level rating cannot capture Wikipedia's within-publication distinctions, such as staff versus contributor bylines at Forbes.

Conclusion

Company articles on Wikipedia are built from a much broader and less prestigious base than the notability guideline's language implies. Technology trade press dominates, led by TechCrunch at 244 citations, which Wikipedia rates no consensus. Of the citations to sources with a rating, 42.3% go to sources the community has not endorsed. Nearly one citation in ten is the company citing itself, which is legitimate for descriptive facts and useless for notability. And the wire release, produced at scale and sold on reach, appears eleven times in two and a half thousand citations. For anyone buying coverage to support encyclopaedic standing, the lesson is that the outlets which actually carry these articles are not the ones tier-one pricing is built around, and that volume of citation is not the same as weight of evidence.

Frequently Asked Questions

Sources

1. vendor research Media Morpheus, September 2026. Original extraction of 2,465 citation URLs from 24 Wikipedia company articles on 24 September 2026, with archived URLs unwrapped to the original publisher and domains cross-referenced against our parse of the perennial sources list. Method and limitations stated above.
2. official Wikipedia. Reliable sources/Perennial sources: the ratings for TechCrunch, Forbes, Business Insider, VentureBeat and PR Newswire referenced above.
3. official Wikipedia. General notability guideline, for the requirement of significant coverage in reliable sources independent of the subject.