Why Wikipedia Pages Get Rejected or Deleted: 12 Failure Patterns from AfC and AfD

The short answer

Wikipedia pages get rejected at Articles for Creation (AfC) or deleted at Articles for Deletion (AfD) for a small, repeating set of reasons — and most of them are diagnosable before anyone writes a word. The dominant cause is notability failure: the subject has not received significant, independent, in-depth coverage, so the article fails the bar even when every fact is true. The next tier is promotional tone and dependence on the subject's own materials (press releases, the company website, interviews). After that come paid-disclosure problems, AI-generated prose, single-event or local-only coverage, awards that carry no weight, founder-and-company conflation, weak draft structure, and post-publication neglect. Of these twelve patterns, roughly half are process-fixable (you can rewrite, restructure, or disclose your way out) and half are evidence-fixable (no rewrite helps until the underlying source coverage exists). This article gives you a forensic framework to tell which is which — so you spend money on the right problem, or correctly decide to wait.

TL;DR

• Most rejections are not about writing quality — they are about evidence. If the independent, in-depth coverage does not exist yet, no amount of polishing saves the draft.
• There are 12 recurring failure patterns. We map each to the exact Wikipedia policy it violates and tell you whether it is fixable by process or only by new evidence.
• A third-party AfC benchmark (Lumino Digital, 2025) reports ~68% of submissions declined, ~16% approval for business topics, and ~6% for startups/tech. Use this as a sober base rate, not a quote of our own data.
• The single highest-leverage move is a source audit before drafting — most failures are predictable from the reference list alone.
• No serious provider can guarantee approval. What a serious provider can do is reduce risk through notability assessment, source research, neutral drafting, transparent paid disclosure, and post-publication monitoring.

The Deletion Forensics Taxonomy

When a draft is declined or an article is nominated for deletion, most people ask the wrong first question: "How do we fix the writing?" The right question is "Which kind of failure is this?" Two failures can look identical on the page — a thin, promotional draft — yet have opposite cures. One needs an editor. The other needs eighteen months of earned media before it should be attempted at all.

The Deletion Forensics Taxonomy is our model for classifying any rejection or deletion along two axes, so you route it to the correct remedy.

Axis 1 — Failure layer. Where did the article actually break?

• Evidence layer — the underlying sources do not support a standalone article (notability, sourcing, coverage depth). No on-wiki action fixes this; only the real world can.
• Representation layer — the sources may exist, but the draft mishandles them: promotional tone, original synthesis, bad structure, citation formatting.
• Process layer — the article may be defensible, but the way it entered Wikipedia is flawed: undisclosed paid editing, COI editing of the article directly, LLM-generated text, post-publication abandonment.

Axis 2 — Fixability. Is the problem process-fixable (a competent, disclosed editor resolves it with materials you already have) or evidence-fixable (nothing changes until new, independent, in-depth coverage exists in the world)?

A failure's coordinates on these two axes tell you the cost, the timeline, and whether to proceed at all. A representation-layer, process-fixable problem is a few days of careful rewriting. An evidence-layer, evidence-fixable problem is a PR-and-time problem masquerading as a Wikipedia problem — and drafting before solving it is how money gets burned.

Every one of the 12 patterns below is tagged with its layer, governing policy, and fixability. That tagging is the taxonomy in use. The official deletion policy exists precisely because pages "that do not meet the relevant criteria for content of the encyclopedia are identified and removed from Wikipedia" (WP:DEL) — so the taxonomy's job is to predict that judgment before it happens to you.

How the three deletion routes differ

"Deletion" is not one process. It is three, with very different speeds, and knowing which you face changes everything.

The distinction matters because the worst outcomes are the fast ones. PROD is gentlest — "any editor… may object to the deletion by simply removing the tag" (WP:PROD), so a single defender cancels it. AfD is a real argument you can win on the merits if the sources are there. Speedy deletion is the one to fear: it is designed to "bypass deletion discussion… and immediately delete" pages, so a promotional draft can vanish before you even see the tag. Patterns 2 and 4 below most often trigger that fast route.

The 12 failure patterns

Each pattern names the layer, the governing policy, and the fixability verdict from the taxonomy. Read the fixability tag first: it tells you whether your next call is to a writer or to your PR team.

1. Weak notability — the coverage is not there yet

Layer: Evidence. Policy: WP:GNG / WP:NCORP. Fixability: evidence-fixable only.

This is the single largest cause of rejection, and the one people most resist hearing. Wikipedia's general notability guideline says a topic is "presumed to be suitable for a stand-alone article… when it has received significant coverage in reliable sources that are independent of the subject" (WP:GNG). For organizations the bar is higher: the guideline states plainly that "only a small percentage of the world's organizations meet the requirements for a Wikipedia article" and that "no company or organization is considered inherently notable" (WP:NCORP).

The trap is that "we are a real, successful company" feels like it should be enough. It is not. Revenue, headcount, and a great product are not notability — coverage about you, by people unconnected to you, is. If that coverage does not exist, the draft is dead on arrival and rewriting changes nothing. This is the textbook evidence-layer failure: the cure is in the world, not on the page.

2. Promotional tone — it reads like marketing

Layer: Representation. Policy: WP:NPOV. Fixability: process-fixable.

Neutrality on Wikipedia is, in the policy's own words, "non-negotiable" and cannot "be superseded by other policies or guidelines, nor by editor consensus" (WP:NPOV). The instruction is blunt: "Avoid stating opinions as facts." Words like leading, innovative, award-winning, world-class, visionary are red flags because they assert a judgment the sources have not made in Wikipedia's voice.

Promotional tone is dangerous twice over: it gets drafts declined at AfC, and sufficiently promotional articles in mainspace are a speedy-deletion candidate. The good news: this is usually a representation-layer, process-fixable problem — if the coverage is genuinely there, a neutral rewrite that describes rather than sells will pass. The mistake is treating tone as the only problem when notability is also weak; fixing prose on an unnotable subject just produces a neutral article that still falls to pattern 1.

3. Primary-source dependency — citing yourself

Layer: Evidence. Policy: WP:RS / WP:NOR. Fixability: evidence-fixable.

A reference list dominated by the company's own website, its press releases, its founder's LinkedIn, sponsored content, and self-serving interviews is the classic tell. Wikipedia requires that "articles should be based on reliable, independent, published sources with a reputation for fact-checking and accuracy" (WP:RS). Press releases and interviews fail the independence test even on a respectable outlet's website, because the words originate from the subject.

The deeper rule is no original research: Wikipedia "articles must not contain original research," defined as "material… for which no reliable source has ever been published" (WP:NOR). You cannot narrate your own company's story onto Wikipedia from internal knowledge. If the only sources are connected to you, the failure is evidence-layer: not a citation problem, a coverage problem.

4. Paid-editing disclosure issues — the trust failure

Layer: Process. Policy: WP:PAID / Wikimedia Terms of Use. Fixability: process-fixable — and mandatory.

This pattern is unique because it is a rules violation regardless of how good the article is. Wikipedia's policy is explicit: "If you were paid or expected to be paid in any way for contributing to Wikipedia, you must disclose it," and "editors who fail to disclose paid contributions are prohibited from editing" (WP:PAID). The binding legal source is the Wikimedia Foundation Terms of Use, which require you to "disclose each and any employer, client, intended beneficiary and affiliation with respect to any contribution for which you receive, or expect to receive, compensation" (Terms of Use §4).

Undisclosed paid editing, when discovered, can get the account blocked and the article deleted on process grounds alone — independent of merit. That is exactly why the "we work confidentially / no one will know it was paid" pitch is a red flag, not a feature. The compliant path is also the safer one: disclose the client on the editor's user page or the article talk page, and — because conflict-of-interest editors "are strongly discouraged from editing affected articles directly" — propose changes "on article talk pages instead" or via Articles for Creation (WP:COI). Done right, disclosure takes minutes. Done wrong, it is the failure no rewrite can undo.

5. AI-generated prose — the new fast-decline

Layer: Process. Policy: WP:AFC / WP:NOR. Fixability: process-fixable, but with an evidence trap.

Reviewers are now fast at spotting machine-written drafts: invented citations, sources that do not say what the text claims, generic encyclopedic filler, and references to publications that do not exist. The AfC process states directly that "articles that are generated entirely by LLMs will be rejected" (WP:AFC).

The most damaging part is not the style — it is the fabricated references. An LLM that hallucinates a Reuters article that was never written hands the reviewer proof of bad-faith sourcing, poisoning the whole draft. There is also an evidence trap: AI prose can make a subject look sourced when it is not, masking a pattern-1 notability failure underneath. Rewriting in a human voice is process-fixable — but if the invented citations were the only thing making the subject look notable, you are back to an evidence-layer problem.

6. No sustained coverage — one good week is not enough

Layer: Evidence. Policy: WP:NCORP / WP:GNG. Fixability: evidence-fixable (time).

A funding round covered by five outlets in the same 48 hours can feel like a win. Reviewers read it differently: a burst of near-simultaneous, often press-release-driven coverage that then goes silent. Notability is better understood as durable notice — has the subject "attracted the notice of reliable sources" repeatedly, over time, in ways that are genuinely independent (WP:NCORP)? The GNG's "significant coverage" must address the topic "directly and in detail," not in a recycled wire story (WP:GNG). This is evidence-layer; the only "fix" is patience — coverage that recurs across months and years. A briefly newsworthy but not durably covered subject should usually wait.

7. Single-event notability — known for exactly one thing

Layer: Evidence. Policy: WP:GNG (and the people guideline, WP:BIO). Fixability: evidence-fixable.

Related to pattern 6 but distinct: the subject is covered substantially, but only in connection with a single incident — one viral campaign, one lawsuit, one controversy. Wikipedia generally resists building a standalone biography or company article around one event, preferring to cover the event itself. For people, the guideline reminds that being "famous or popular… is secondary" to having "significant coverage in reliable secondary sources that are independent of the subject" (WP:BIO). The fix is not editorial — it is more, and more varied, independent coverage establishing the subject matters beyond the one event.

8. Awards and rankings that do not count

Layer: Evidence. Policy: WP:RS / WP:NCORP. Fixability: evidence-fixable.

"We won three industry awards" is one of the most common notability arguments — and one of the weakest, because most awards are pay-to-enter, vendor-run, or "best of [city]" lists with no independent editorial selection. Wikipedia weighs sources by their "reputation for fact-checking and accuracy" (WP:RS); a directory that lists anyone who applies or pays carries no such reputation. A genuinely selective, independently judged honor with substantial secondary coverage about the win can help. A logo on your homepage cannot. This is evidence-layer: you cannot rewrite a weak award into a strong one — only seek the independent coverage that moves the needle.

9. Local-only coverage — the geography trap

Layer: Evidence. Policy: WP:NCORP. Fixability: evidence-fixable.

Routine coverage in a single local outlet — the hometown business journal running a profile of every new shop — rarely establishes notability for a company. Wikipedia bases the decision on whether the organization "has attracted the notice of reliable sources unrelated to the organization" in a way that is significant, not on volume of routine, local, announcement-style mentions (WP:NCORP).

This is especially relevant for European businesses: strong coverage may exist in national or regional press in a non-English language. That coverage does count — Wikipedia accepts non-English reliable sources — but it must be genuinely independent and substantial, not the local-paper-prints-the-press-release variety. (For which language edition to attempt first, see our multilingual Wikipedia strategy guide.)

10. Founder-and-company conflation — notability does not transfer

Layer: Evidence. Policy: WP:BIO vs WP:NCORP. Fixability: evidence-fixable.

A frequent failure: a draft about a founder that is really about their company, or vice versa. The two are judged under different guidelines and notability does not transfer between them. A notable company does not make its CEO notable, nor does a notable founder automatically make their startup notable. The people guideline applies its own test — "significant coverage in reliable secondary sources that are independent of the subject" — to the person (WP:BIO), while the organizations guideline insists "no company or organization is considered inherently notable" for the company (WP:NCORP). Decide which subject you are arguing for, and assemble independent sources about that subject. Our deep dive on this split is Companies vs Founders vs Public Figures.

11. Poor draft structure — defensible content, indefensible packaging

Layer: Representation. Policy: WP:V / WP:NPOV. Fixability: process-fixable.

Sometimes the subject is notable and the sources exist, but the draft is built badly: claims without inline citations, a wall of unsourced "About us" prose, missing lead, no structure a reviewer can verify. Wikipedia is explicit that "the burden to demonstrate verifiability lies with the editor who adds or restores material," satisfied "by providing one inline citation to a reliable source" (WP:V). A reviewer should not have to hunt for what supports a sentence. This is the most genuinely process-fixable pattern of all — same facts, same sources, correct packaging — and the cheapest failure to fix or prevent.

12. Post-publication neglect — surviving AfC is not the finish line

Layer: Process. Policy: WP:DEL / WP:AFD. Fixability: process-fixable (ongoing).

A published article is not permanent. It can be nominated for deletion years later, edited by anyone, vandalized, or quietly degraded by a competitor adding negative framing. AfD discussions run "for at least seven days" before a consensus close (WP:AFD) — and if no one who understands the sourcing defends the article in that window, it can be lost for entirely answerable reasons. Deletion "removes the current version and all previous versions from public view" (WP:DEL), so neglect can erase years of standing. This is process-fixable but ongoing: monitoring, a defensible source base kept current, and someone ready to engage a deletion discussion on the merits. We cover the full lifecycle in After Publication: monitoring and the 5-year risk curve.

The failure-pattern matrix (the framework instrument)

This is the Deletion Forensics Taxonomy applied to all 12 patterns at once. Find your symptom, read across to the fixability column, and you will know whether your next step is a rewrite or a reckoning with the evidence.

Read the pattern in the column. Eight of the twelve are evidence-fixable. That is the uncomfortable, useful truth of this whole article: the majority of Wikipedia failures are not writing problems. They are evidence problems wearing a writing problem's clothes — which is why a source audit, not a rewrite, is the correct first spend.

A soft note before you go further. If your honest read of the matrix is "most of my problems are in the evidence column," the right move may be to not attempt a page yet — and to invest in the coverage that makes one defensible later. For a second pair of eyes on which column you are actually in, our Notability Audit answers exactly that, before any drafting money is spent.

What the base rates actually look like

Calibrate expectations against real, published numbers — attributed, not invented. The most-cited public AfC dataset comes from Lumino Digital, whose 2025 analysis of just over a thousand Articles-for-Creation submissions reported an overall decline rate of roughly 68%, an approval rate near 16% for business-category topics, and about 6% for startups and tech companies (Lumino Digital, AfC approval-rate report, 2025). We cite this as a credible third-party benchmark; it is not WikiBusines's own measurement.

Two caveats. These are AfC submission rates, skewed by the volume of clearly-unnotable attempts — a pre-screened, genuinely notable subject faces far better odds than the raw 16% suggests. And no public dataset is Europe-specific, so treat the figures as a direction, not a guarantee. The lesson stands regardless of the exact percentage: business and startup topics are the hardest categories on Wikipedia, and the failures cluster in the evidence column above.

A decision tool you can use right now: the Source Strength Pre-Mortem

Before contacting anyone — us included — run your own reference list through this checklist. Score each prospective source. You can do this in a spreadsheet in twenty minutes, and it predicts most AfC outcomes better than reading the draft does.

Score each source you intend to cite, 0–3 on four axes:

Now total each source and apply the rule of thumb:

• A source scoring 10–12 is a load-bearing source. You typically need three or more of these, about genuinely different events or angles, for a defensible company article.
• Sources scoring 4–9 are supporting — useful for facts, not for carrying notability.
• Sources scoring 0–3 are decorative — they will not help notability and, if they dominate your list, they signal patterns 1, 3, 8, or 9.

Decision: If you cannot find at least three load-bearing (10–12) sources that are independent, in-depth, and not clustered around a single event, your draft will probably fail pattern 1 — and the correct action is to build coverage first, not draft now. If you have them, your risk shifts to the process column, where careful drafting and disclosure do the work.

This is the logic our Notability Audit formalizes — but the checklist above is yours to run for free, and you should, whether or not you hire anyone. (The downloadable Source Risk Audit Sheet below turns it into a ready-made grid.)

What we will NOT promise — and why

We will not promise approval. Not "guaranteed approval," not "100% acceptance," not "your page won't be deleted." Anyone who promises a Wikipedia outcome is either misunderstanding how Wikipedia works or asking you to ignore how it works. Wikipedia is governed by a volunteer community and a body of policy; no vendor controls reviewers, administrators, or consensus, and no honest one claims to. The general notability guideline itself says significant coverage "creates an assumption, not a guarantee, that a subject merits its own article" (WP:GNG). What a serious provider can do is reduce risk before money is spent — through an honest notability assessment, real source research, neutral drafting, transparent paid disclosure under WP:PAID, and post-publication monitoring. When we are confident, we publish our confidence terms on /guarantees: if a published page cannot be defended after three attempts within the 90-day monitoring window, we refund 80%. That is a statement about our work and our risk-screening discipline — not a claim to control the community. The difference is the whole ethics of this field.

Pricing and where the money should actually go

Because most failures live in the evidence column, the highest-value early spend is diagnosis, not drafting. Our Notability Audit is EUR 490 / 750 / 1,900 (roughly USD 530 / 815 / 2,065) depending on depth and the number of subjects or languages assessed — and the fee is credited toward the project if you proceed. It exists to put your reference list in the matrix above before you commit to creation.

If a subject is genuinely defensible, creation pricing is anchored as follows (per page; approximate USD alongside; figures depend on source strength, language edition, complexity, COI sensitivity, and maintenance):

USD figures are approximate and move with the exchange rate. A full breakdown of one-time versus recurring cost — including the post-publication monitoring that defends against pattern 12 — lives in our Wikipedia page cost guide and pricing guide. The honest framing: spending EUR 1,930 to draft an article whose sources score in the decorative band is how money gets deleted, not just spent. Audit first.

Frequently asked questions

Why do Wikipedia pages get deleted? Almost always for one of the 12 patterns above, with notability failure (no significant, independent, in-depth coverage) as the dominant cause. Promotional tone, dependence on the subject's own sources, and undisclosed paid editing are the next most common. The official basis is WP:DEL, which removes pages that "do not meet the relevant criteria for content of the encyclopedia."

What are the most common reasons for Wikipedia article deletion? In order of frequency: weak notability, promotional/non-neutral tone, poor or primary sourcing, single-event or non-sustained coverage, and process problems like undisclosed paid editing or LLM-generated text. Copyright violation and blatant promotion are the fastest routes, because they qualify for speedy deletion under WP:CSD. Most others are argued over seven days at AfD.

Can a deleted Wikipedia page be restored? Sometimes — depending on why it was deleted. A page removed for fixable reasons can be improved and resubmitted, or taken to deletion review; a page deleted for genuine non-notability cannot return until real coverage exists. We walk through the options in How to get a deleted page back and our recovery service.

What is speedy deletion on Wikipedia? It is the fastest route: a single administrator may delete a page "without waiting for any discussion" when it meets a defined criterion such as blatant promotion or copyright violation (WP:CSD). It is the outcome to fear most, because it can happen within hours and is what overtly marketing-style drafts trigger.

How long does a Wikipedia deletion discussion last? An AfD discussion runs "at least seven days," after which an administrator closes it based on community consensus (WP:AFD). Proposed deletion (PROD) also runs seven days but is cancelled the moment any editor removes the tag (WP:PROD).

Can promotional language get my page deleted? Yes. Neutrality is "non-negotiable" on Wikipedia (WP:NPOV), and sufficiently promotional articles are a speedy-deletion candidate. If the underlying sources are strong, a neutral rewrite usually saves it; if they are weak, fixing the tone alone will not.

Who decides if a Wikipedia article is deleted? For AfD, the community discusses and an uninvolved administrator closes on consensus; for speedy deletion, a single administrator decides; for PROD, an admin acts only if no one objects. No external party — and no agency — decides or controls this, which is why outcomes cannot be guaranteed.

Why was my company's draft rejected at AfC? Most often pattern 1 (notability) or pattern 3 (primary sourcing) — the reviewer did not find enough independent, in-depth coverage to justify a standalone article. The fix depends on the column: if it is evidence, you need more coverage; if it is representation, you need a better-sourced, neutral rewrite. Our companion piece Why your Wikipedia draft was rejected breaks down the decline reasons.

How do I save my page from deletion? Show up to the discussion with the sources. In an AfD, the article is kept or deleted on the strength of independent, significant, reliable coverage demonstrated during the seven days — so the practical defense is a strong source base and someone who can argue it on the merits, which is what monitoring provides.

Can promotional tone really be the only problem? Occasionally — a notable subject packaged badly (pattern 11) is genuinely process-fixable. But treat "it's just the tone" with suspicion: very often weak tone and weak notability travel together, and fixing only the prose leaves a neutral article that still fails pattern 1.

Get the Source Risk Audit Sheet

Score every citation before a reviewer does. The Source Risk Audit Sheet turns the Source Strength Pre-Mortem above into a ready-to-use scoring grid: paste your references, rate each on independence, depth, reliability, and secondary-vs-announcement, and the sheet flags your load-bearing, supporting, and decorative sources — plus a clear go / wait verdict against the 12-pattern matrix. It is free, and the single most useful thing you can do before spending on a draft.

Form fields:

• Work email (required) — where we send the sheet
• Name (required)
• Subject of the page (required) — single line (company or person name)
• Subject type (required, select) — Company · Founder/Executive · Public figure · Product/Brand · Other
• Primary target language edition (optional, select) — English · German · French · Spanish · Ukrainian · Other
• Roughly how many independent sources do you have? (optional, select) — 0–2 · 3–5 · 6–10 · 10+
• Consent (required, checkbox) — "Send me the Source Risk Audit Sheet and occasional Wikipedia-risk guidance. I can unsubscribe anytime."

Submit button: Send me the Audit Sheet. No sales call is required to receive it.

The complete 2026 Wikipedia playbook

This guide is one part of a ten-part series — an honest, end-to-end walkthrough of getting and keeping a Wikipedia page in 2026. Each part stands alone; together they cover the whole journey.

Before you start — Can my company get a page? · Company vs founder vs public figure Budget & vendor — What it costs — 5-year TCO · The honest vendor scorecard Compliance & risk — Paid editing, COI & disclosure · Why pages get deleted — 12 patterns (you are here) Strategy & growth — Wikipedia, Wikidata & AI search · Multilingual strategy After publication — Monitoring & the lifecycle risk curve The data — Wikipedia Risk Report 2026

Not sure where your case stands? A fixed-scope Notability Audit reads your real sources against policy — or just talk to the team.