What do the three core policies actually require?
They require that every claim be verifiable in a reliable published source, that editors add no analysis the sources did not make, and that sources meet the reliability guideline. Together, Verifiability, No original research and Reliable sources form the base every reviewer applies silently.
WP:V sets the threshold: material must be checkable against published sources, whether or not it is true in some absolute sense. A founder's "we really invented it first" is invisible to the encyclopedia until a reliable source has said so. WP:NOR stops editors from connecting dots the sources did not connect. WP:RS defines which sources qualify, on three tests that must all hold for the claim in question:
- Independence. The source is not owned, controlled or paid by the subject and is not a rewrite of its material. A vendor-owned trade site, a "media partner" carrying client content and a feature lifted from a press release all fail.
- Editorial oversight. Someone assigns, edits, fact-checks and issues corrections. A byline and a masthead are not enough; what matters is whether an editor stood between submission and publication.
- Reputation for accuracy. The outlet has a track record, recognised by other reliable outlets. Young outlets often sit in "no consensus" because the record has not accumulated yet, not because anything is wrong.
The practical consequence: you do not persuade anyone that you matter. You show that reliable sources have already covered the fact that you matter. The encyclopedia follows the press, not the reverse.
How do Wikipedia editors rank source types?
There is no official table, but experienced reviewers carry a similar four-band map. Treat the bands below as our working summary of review practice, not as Wikipedia policy; official statuses live on the Perennial Sources list (WP:RSP).
| Band | Typical sources | How reviewers treat them |
|---|---|---|
| 1. Settled | International wires and newspapers of record (Reuters, AP, BBC, Financial Times, The Guardian), major scientific journals | Default assumption: good citation, unless the piece itself is opinion, sponsored or a relayed release |
| 2. Reliable, narrower weight | National broadsheets outside the English-speaking world, public broadcasters, strong specialist magazines | Count for most claims; judged on WP:RS where no RSP entry exists |
| 3. Situational | Regional dailies, trade press, university presses, outlets rated "no consensus" | Count together, not alone; trade press leans toward boosterism and wants a more critical outlet beside it |
| 4. Does not count | Press-release wires, contributor networks, corporate blogs, sponsored content, pay-to-publish "as featured in" sites, self-published platforms | Weak or damaging: a list padded with them signals a thin source base |
Our media list shows why the bands cannot be applied by logo alone. PR Newswire is rated generally unreliable on RSP because its releases carry no editorial oversight, and Medium is rated generally unreliable because it is a hosting platform rather than an outlet. Contributor columns on business magazines are rated generally unreliable and treated as self-published, while the same magazines' staff writing is judged case by case. The byline decides, not the logo; our Forbes and Daily Mail guide walks through that split.
Which source types are edge cases?
Most edge cases come down to one question: is the source primary, self-published or opinion, and what claim is it being asked to support? These come up most often.
- Preprints (arXiv, SSRN, bioRxiv). Generally not reliable, because peer review is the editorial process WP:RS looks for. Exceptions exist for established authors on established results, but a draft built on preprints will be challenged.
- Conference proceedings. Field-dependent. In computer science, selective peer-reviewed conferences are often the main publication venue and are treated close to journals. In medicine and most social sciences, abstracts carry far less weight.
- Opinion and editorial. Reliable for the opinion of the writer or the editorial board, attributed as such. A flattering op-ed about your CEO supports "X argued that…", not a fact claim inside the argument, and weighs little for notability.
- Self-published expert sources. WP:SPS allows a narrow exception only for an established subject-matter expert whose work has been published by independent reliable sources. Expertise has to be demonstrated externally first.
- Industry analyst reports. Usable for a discrete attributed fact ("ranked as a leader in the 2025 report on…"), weak for notability, since one analyst is one view.
- Court records and filings. Primary. They can support a numeric fact; the framing of why it matters must come from secondary coverage of the filing.
- Social media. Not reliable for third-party claims. Narrow use as evidence of what an official account itself stated, if uncontentious. Never notability.
- Press releases and company announcements. Never count toward notability. Usable for plain, uncontroversial facts such as founding year or headquarters. A news outlet that reprints a release inherits its primary character.
- Databases and directories. Crunchbase is deprecated on RSP, and similar listings show you exist without showing significance, since they list everyone.
How do reviewers score a source list?
They look for the strongest two or three sources and ask whether those alone make the case. They do not add up a column of citations. Articles for Creation (AfC) is the volunteer review queue for new drafts, and its reviewers apply conventions that are not in any single policy page.
Significant coverage.
The unwritten test is whether you can write several encyclopedic sentences from the source alone. A name in a list of attendees, or a quote in a roundup, is trivial coverage however reliable the outlet.
Independent and secondary.
The organizations guideline (WP:NCORP, including WP:ORGIND) excludes content built on company-supplied information, interviews where the subject is the only voice, and announcements relayed without reporting. Routine news such as funding rounds, hires, launches and quarterly results does not establish notability on its own, because every company generates it.
The three-sources habit.
In practice reviewers look for around three independent, reliable, secondary sources with significant coverage. That is a convention, not a rule: Wikipedia sets no number. One reported feature in a top outlet can outweigh several regional mentions, while piling up weak citations cannot rescue a thin base.
Time spread.
A single news cycle years ago reads as a spike. Several independent pieces spread across months read as ongoing interest, which is what notability is meant to capture. A sudden recent surge with no history often ends in "wait and see".
The syndication trap.
One wire story reprinted on fifty sites is one source. If two pieces share most of their paragraphs, they are one piece.
The result can be sobering: a company with twenty press hits may have two or three citations of substance once reprints, contributor placements, routine announcements and passing mentions are removed. Reviewers do that stripping by reflex; do it first.
How do you turn a press list into a usable source set?
Inventory, de-duplicate, rate each piece on four axes, then sort into pass, borderline and fail. Run it before any drafting.
- Inventory. Every meaningful piece from recent years: URL, outlet, byline, date.
- Collapse reprints. Merge every copy of one story into one entry, translated reprints included.
- Rate each survivor. Is the outlet reliable (check RSP and the byline)? Independent? Secondary? Significant, meaning about you in depth?
- Sort. Pass clears all four. Borderline clears most and will be argued over. Fail does not count.
- Read the result honestly. Our rule of thumb, not Wikipedia's: several independent passes over time make a page realistic; a few passes among many borderlines is uncertain; all borderlines and fails means "not yet".
For orientation by sector, our notability index gives our own assessment of how often companies in 13 industries clear the bar, by year from 2018 to 2026. For outlet-by-outlet lookups use the media list. If the sources do not exist, the fix is an earned-media project, not a Wikipedia project, and a Wikidata entry has a lower bar while that work happens.
Does the same source layer matter for AI answers?
Yes, though nobody can promise what an AI will say. Answer engines retrieve from the open web, and independent, reliable, secondary sources are the material they can verify. Wikipedia is also a heavily used source, so a well-sourced page can shape how a brand is described. A footprint made of press releases and contributor columns gives them thin or contradictory material. See our AI visibility work.
How WikiBusines helps
We run this assessment on your real coverage and say plainly when the answer is "not yet".
- Notability audit, from €490, credited toward the page if ordered within one month: a written verdict with reasoning per source.
- Page creation, from €2,400 for English, 4 to 5 weeks, with 90 days of monitoring and guarantee after publication. Paid editing is always disclosed on Wikipedia in WikiBusines' name. If a published page cannot be defended after 3 attempts within the 90-day window, 80% is refunded.
- Annual support, from €720 per year: up to four updates, with a 12-month guarantee.
We never guarantee that an article will be published. Start with the free 10-minute assessment: contact us. If the case is unclear, we offer the paid notability audit from €490, credited toward the page if you order it within one month.
Frequently asked questions
Independence from the subject, real editorial oversight and a reputation for accuracy, judged per claim under WP:RS. Community consensus on recurring outlets is kept on the Perennial Sources list.
On the live Perennial Sources list, which can change. Our 2026 statuses guide summarises it, and the media list sorts outlets by suitability.
The test is verifiability, not truth. A claim that no reliable source has published cannot be added, however accurate it is.
Only for plain, uncontroversial facts about the company, such as a founding year. They are primary and not independent, so they never establish notability.
For basic facts about itself, with limits. It cannot support contested or promotional claims and carries no weight for notability.
There is no fixed number. Reviewers look for significant coverage in multiple independent reliable secondary sources, and in practice about three strong ones is a common reference point.
Generally not, since they have not passed peer review. Established authors on established results are the narrow exception.
No. Paid placement is not independent, even in a reputable outlet. Treat it as promotional material.
Authors and editorial policy
Bohdan DubylovskyiCo-founder and CBO, WikiBusines→
Roman MelnykCo-founder & Chief Editor, WikiBusines→The article was prepared by the WikiBusines co-founders. We sell paid Wikipedia services, so we disclose this interest: paid editing is always declared on Wikipedia in WikiBusines' name, and claims about the rules are backed by links to primary sources.
Editorial policy: payment disclosure, conflicts of interest, sources