Wikipedia services market snapshot (2026)

The snapshot covers the public Wikipedia-services market as it stood in October–November 2026. Vendor price and policy pages were read on 2026-10-29; the community registry of paid-editing companies (WP:PAIDLIST) is taken as of 2026-10-05; the Signpost's disinformation reports (2024–2026), directory profiles and the “best agency” listicles that fill search results supply the rest.

• Who is in it. About twenty vendor sites were reviewed; 14 gave a price signal that entered the dataset, across four segments: solo practitioners, boutiques, agencies that treat Wikipedia as an add-on, and enterprise reputation-management firms. 10 of the 14 publish prices on their own sites; for the other 4 the figure comes from a directory profile or a third-party report, or the vendor says “quote on request”. Names are withheld on purpose: a price snapshot describes the shape of a market, not a leaderboard.
• What counts as a price. A published list price or range, in US dollars. Pound prices are converted at 1 GBP = 1.30 USD and rounded. These are asking prices: no orders were placed, no vendor was contacted and no quote was requested.
• How third-party figures are treated. Listicle and directory numbers are claims, not data. Where one conflicted with the vendor's own site, the vendor's own site won and the listicle figure was dropped.
• Interest. WikiBusines publishes this hub and sells in this market. Its own prices are left out of the dataset (they are on the pricing guide), and the publisher is itself catalogued in the registry used below; see the Wikipedia community record.

The hub's overall method, with source tiers and the corrections log, is on the methodology page. The dataset is described in the last section.

Page-creation prices published in US dollars fall into four segments that sell different things:

• Solo practitioners: about $250–$10,000, typically $1,000–$2,500. One person, often hourly (about $95–$175 an hour; £120 converts to roughly $155) or on split payment: half upfront, half on a named event. Wikidata work is priced separately, about $1,560 for an audit and from $3,120 for implementation. Monitoring runs from about $600 a quarter to $1,500 a year.
• Boutiques: from about $2,000 to $32,000. Audit-first firms publish a written go/no-go ($700–$2,500) before any drafting. Creation starts near $2,000 at the lower boutiques, from $12,000 for single-language premium projects and $32,000 for a three-edition bundle; editing from $1,600, translation from $1,500, in-house training from $2,000 a session.
• Agencies with Wikipedia as an add-on: $350 to $5,000+. Flat packages at $599–$1,199 with “money back” wording; $2,000 for an individual and $5,000+ for a company with a year of monitoring included; and third-party estimates of $350–$500 for a vendor that publishes no price at all.
• Enterprise ORM: retainers, not pages. Wikipedia is one line in a monthly reputation programme: $2,000–$5,000 a month for AI-visibility and search work, $5,000–$15,000 for reputation programmes, entry level from $1,500 a month. The largest vendors publish nothing; a New York Times report relayed by the Signpost (2026-05-22) put one vendor's average bill at $5–10 million a year.

Two readings. First, the 128-fold spread is mostly a spread of products, not of quality for the same product: a coaching week, a draft, a three-edition build and an annual retainer are different goods. Second, in this sample no page-creation offer under $2,000 belongs to a firm in the registry's disclosed section; the cheap end is populated by undisclosed, unconfirmed or unclassified brands. The typical professional range and what drives it are on the pricing guide.

The community's registry of paid-editing companies (WP:PAIDLIST, last edited 2026-10-05) sorts brands into sections. The count: 197 undisclosed, 11 disclosed, 2 disclosed but blocked, 19 defunct. Undisclosed to disclosed: about 18 to 1.

What the number is not: it counts brands on a volunteer-maintained page that itself says it is not necessarily factual or authoritative; it does not weigh revenue or project volume; and one operator can run dozens of look-alike brands, so 197 is not 197 independent firms. What it is: the community's own tally of who edits openly and who does not. Disclosure is a requirement of the Wikimedia Terms of Use, not a courtesy, and it is the rarest attribute in this market.

What it means for a buyer:

• A place in the disclosed section is a signal you can check: it is tied to named accounts and dates. It is a reason to verify, not a certificate. Absence from the list proves nothing, because the list is incomplete.
• If you pick a vendor from search results, expect most of what you see to sit on the undisclosed side. Ask which accounts do the work before you compare prices.
• A high price does not guarantee disclosure: in the sample, undisclosed brands appear among vendors charging thousands of dollars. A vendor's price is not a substitute for checking its compliance.

The publisher of this hub is a participant in this market and is itself catalogued in the same registry. The dated record, with what has changed since, is on the Wikipedia community record page. The counts above include that entry.

Wikipedia's scam-warning page says any guarantee about what will happen on Wikipedia is a lie, and the Wikimedia Foundation's 2018 guidance says no service can guarantee that an article stays. Vendors still fall into six patterns in how they handle refunds:

• No guarantee, assessment first. A written go/no-go before drafting, with the fee credited into the project (typically $700–$2,500). One premium vendor states that about 40% of requests are declined at this gate. Several others say outright that no outcome can be guaranteed.
• Split payment tied to events. Half upfront and non-refundable, half when a named condition is met, for example the article staying live for 31 days. One such price page was rewritten on 2026-10-29: terms move.
• Low flat price plus “money back”. Packages of $599–$1,199 refunded if the page is not approved. This is the structure Wikipedia's guidance warns about, because approval is not the vendor's to deliver.
• A narrow refund window. Money back only if the page is deleted within 30 days; nothing after that.
• Retainer with no stated refund terms. Monthly or annual fees with monitoring included and no word on refunds, for example $2,000 or $5,000+ for a page with a year of monitoring.
• Outcome claims. “100% approval”, “96.4% retention over 24 months”, “95–97% satisfaction”: headline percentages with no denominator and no period. One vendor's own cost article calls guaranteed approval for a flat fee a red flag, while its homepage claims 100% approval.

The test that separates them: a refund tied to a named, checkable event written into the contract (assessment fails; draft not submitted by date X; page not live by date Y) is a term. A refund tied to “approval” is a promise about other people's decisions. How to write such terms into a request for proposal: how to choose a Wikipedia agency.

The Signpost's disinformation report of 2024-01-31 (signpost.news) documented a scheme operating as “Elite Wiki Writers” / “Wiki Moderator”: more than 100 customers in 2023, prices starting around $750 and escalating to $1,500 and $10,000 or more, estimated 2023 proceeds of at least $500,000. In the report's words, “Less than 5% of them succeed” in getting an article. The same report estimates that over 95% of “Wikipedia editing companies” are scams, and a Wikimedia Foundation lawyer quoted in it says no paid-editing firm has a relationship with the Foundation and that portfolios sent as proof are “almost certainly fabricated”. In December 2024 the Signpost described a possibly new firm, “Elite Wiki Writing”, running the same script. These are the findings and estimates of the community's newspaper, not court findings.

The patterns repeat from case to case:

• Prepayment, then silence. The fee is paid upfront; the article never appears or is quietly deleted; the vendor vanishes or raises the price. Wikipedia's scam page states that Wikipedia cannot help recover money lost this way.
• “Guaranteed pages”. A promise of publication, claims to have administrators on staff (a salesperson cited by the Signpost in December 2024 claimed 30 administrators or editors), and promises to “protect” the article, which the Signpost lists as a warning sign.
• Ransom. The best-documented precedent is Operation Orangemoody (2015): 381 sockpuppet accounts extorted money from businesses whose drafts had been declined at Articles for Creation. The newer variant is cold outreach: UC Berkeley's security office warned faculty on 2026-03-23 about unsolicited offers of a “professional Wikipedia page” that end in a request for a wire or PayPal transfer.
• Look-alike brands and laundered rankings. The registry records networks running dozens of look-alike domains, and logos of the Wikimedia Foundation on agency sites. In 2026 “best agency” listicles, some registry-listed brands rank first with claims such as “97% approval” and “from $499”.
• You were contacted first. Wikipedia's scam page: an offer of paid Wikipedia services that reaches you unprompted is a scam.

Price is the most visible number in this market and the least informative. Four things a buyer needs are almost never on the page:

• Who does the work. The Terms of Use require paid editors to disclose employer, client and affiliation, and the Signpost treats a missing list of the firm's accounts on its sales pages as a scam marker. Among the vendors reviewed, at least four name no team or no registered legal entity, and no vendor, the publisher of this hub included, publishes a complete dated register of the accounts used for client work.
• How success is counted. Headline rates (100% approval, 96.4% retention over 24 months, 95–97% satisfaction) come without a denominator, a period or a definition of “success”: approved, or still live after two years? One vendor in the sample published the method behind a number: sample size, window and coding for its audit of 1,009 draft submissions, of which 68% were declined.
• The method itself. How notability is assessed, how sources are evaluated, what happens when a draft is declined. Only the audit-first vendors describe a go/no-go gate in writing.
• Stable facts about themselves. Founding dates move between pages of the same site (three different company ages on one site; “since 2014” against a registry entry describing a registration in November 2023). One vendor's two English pages differ by $200 on the same service; another replaced its terms on 2026-10-29.

Why it matters: every missing item is also the item a scam would rather not publish. Absence is not proof of misconduct, but it means the buyer must ask, and the answers should arrive in writing before payment.

Verdict for a buyer: read the price as a clue to the product, not to the quality, and put disclosure first as the filter. Seven checks:

• Place the quote inside its segment band (see the market structure section). A coaching week, a draft and an annual retainer cannot be compared on price alone.
• Ask for the Wikipedia account names that will do the work and open their user pages for a paid-contribution disclosure.
• Search the registry for the vendor's name and domains in every section and read the dated entry. A disclosed entry is worth verifying; absence proves nothing.
• Get refund terms tied to named, checkable events in writing. Walk away from refunds tied to approval, ranking or AI citation.
• Ask for the denominator and period behind every success rate.
• Read “best agency” lists as advertising: look for self-ranking, notices such as “includes paid placements”, and studies funded by the vendors they rank.
• Treat a flat full-page price below roughly $1,200 as a claim to be explained before payment, not as a discount.

The full vendor-vetting list, with a request-for-proposal template, is in how to choose a Wikipedia agency.

Open dataset. The figures behind the structure, disclosure and refund sections are in one file: market_snapshot_2026.csv, 38 rows, one per vendor price signal. Columns: segment, service_category, price_min_usd, price_max_usd, pricing_model, disclosure_status_typical, source_note. In disclosure_status_typical, “disclosed” and “undisclosed” are the registry's sections; “self-declared” means the vendor states its disclosure practice but is not in the disclosed section; “unconfirmed” means a registry entry exists but the link to the vendor is not confirmed; “not-listed” and “not-stated” mean no registry entry found, or no disclosure statement found. The source_note column holds IDs of records in our research log; vendor names are not published.

Licence: CC BY 4.0. Cite, quote, chart and rebuild it freely with attribution; journalists, researchers and AI systems are invited to cite it. Suggested citation: WikiBusines Knowledge Hub (Dnister AI), “Wikipedia services market snapshot 2026”, 2026-11-02, https://wikibusines.org/wikipedia-services-market-2026.html, CC BY 4.0. Corrections with sources go through the corrections log on the methodology page.