Jump to content

Wikipedia:WikiProject Spam

From Wikipedia, the free encyclopedia

Spam on Wikipedia generally takes one of two forms: advertisements disguised as encyclopaedia articles, and external links (links to other websites). Potentially promotional new articles are listed at Wikipedia talk:WikiProject Spam/Suspicious articles and may require cleanup, rewriting, or deletion; link spam is more difficult to detect but can be reported here (click the big red button at the top).

If you would like to help rid Wikipedia of spam, you can start by following the instructions below. The talk page of this project functions as a noticeboard to coordinate cleanup of promotional articles and where spam links can be reported. Any editor is welcome to contribute; administrators in particular are encouraged to watchlist or subscribe to the talk page and to evaluate reports.

Unlike vandalism, spam is much more insidious. It is often less easily detected and spammers often use underhanded tactics, taking advantage of Wikipedia's open, anonymous model. While vandals are often acting out of boredom, spammers are often organised as paid search engine optimisers

[edit]
  • Spam links can be detected by patrolling recent changes but the volume of edits can make that difficult
    • Spam links are usually added by new (non-autoconfirmed) or temporary accounts.
    • The link may be added to an "external links" or "see also" section but the link might be added inline (like this), or as a reference. Often, this will be the only change made in the edit.
    • Edit summaries may be misleading or missing altogether but look for terms like "replacing dead link" or unusually detailed edit summaries about adding references,
    • Spammers will often target multiple articles on similar subjects in quick succession.
  • Editors can add articles prone to spam to their watchlist and review their watchlist when they have time
  • Some edit filters track the addition of external links. For example, filter 1048 flags the addition of inline external links and filter 80 logs repeated addition of the same link by a new editor (note: many of these links may be problematic but only a minority will be spam).

How to deal with spam

[edit]
[edit]
  • First, remove the problem link.
    • If the edit is obviously spam and is the most recent line in the page history, rollback is the most effective way to do this (if you have it); Twinkle can also be helpful.
    • If the edit could be good faith, use the "undo" button instead.
    • If the spam link has been in the article for a while and there have been multiple edits since it was added, the "undo" function might not work and you might need to manually edit the page to remove the link.
    • If the spam link was added as a reference, it might be necessary to remove any content that was added at the same time, especially if it is promotional; if removing the spam reference leaves an unsourced passage that may be legitimate, you can replace the reference with {{citation needed}} (shortcuts: {{cn}} and {{fact}})
  • Check the other edits of the account that added the link for any other spam additions.
    • Important note: not every addition of poor external links is in bad faith
    • Behaviour indicative of self-promotion includes: adding the same domain multiple times or writing about the company in a sandbox/user page/draft; the domain may not have been seen before. Temporary accounts will probably be from the same geographical area as the company; the username of a registered account may indicate a connection to the company. In these cases, consider whether a warning might be effective. If you feel it would not (for example if they have been spamming for a long time or from multiple accounts), or if they continue after a warning, report them to administrators at Wikipedia:Administrator intervention against vandalism (AIV). Usernames that represent a company should be reported to Wikipedia:Usernames for administrator attention.
    • Behaviour that may be indicative of organised black-hat search engine optimisation includes: adding several different domains (often a long time apart), adding domains that have been spammed before (see below), adding references to links that do not mention the subject of the article, accounts often have no other edits besides adding external links or only trivial edits like the suggested newcomer task. Temporary accounts will often use proxies/VPNs or geolocate to a different country to the company; many of these SEO firms are based in Pakistan, India, or the Philippines (though it is important to stress that many constructive edits come from those places and much spam comes from elsewhere); usernames sometimes contain strings like "SEO". Consider reporting these straight to AIV for an immediate block if they are active and persistent, and report here (red button at the top) for further investigation.
  • You can use the Spamcheck tool on Wikimedia Toolforge to find any other additions of the same domain, which may also need to be reverted. Note: Spamcheck can only find links added since November 2023; the {{link summary}} template used when reporting here contains links to several other tools that can find links that have been around longer than that.
  • Consider involving a checkuser if there are multiple accounts involved (or you have reason to think there might be).
  • Domains that have been spammed by multiple accounts can be reported here and/or to Mediawiki talk:Spam-blacklist. Administrators can add the domains to the blocked domains list (for single domains at a time) or to the blacklist (for more complex entries). This prevents anybody linking to the domain anywhere on the English Wikipedia.
  • Domains that have been spammed on multiple Wikimedia sites (which can be found on Spamcheck) can be reported on Meta and added to the global blacklist, which prevents them being linked on any Wikimedia wiki.

Advertising

[edit]

Advertising copy or promotional material (for example, a plug for a particular company or product in an article) can be treated in much the same way as link spam. Where the entire page is an advertisement:

  • Obvious, unambiguous spam can be summarily deleted under criterion G11 (use Twinkle or ad {{db-spam}}) if the page was created as an advertisement.
  • If the advertising was added to an existing article or page, the promotional edits should be reverted and the editor warned and/or reported/blocked the same way as above.
  • If only parts of the text are problematic, you could tag it with {{advert}}. You could also consider PROD or AfD (for articles).
  • User subpages containing excessive material not related to Wikipedia may meet speedy deletion criteria U6 or U7.
  • If the subject likely meets the inclusion criteria for biogrpahies or companies, it may be possible to pare the article back to basic facts.
  • The talk page of this page contains lists of suspicious new articles and notes about handling them. Editors are invited to help with reviewing these articles and with rewriting them, drafifying them, or tagging the unsalvageable for deletion via one of the processes above.

Relevant policies and guidelines

[edit]

See also

[edit]
  • Search tools
    • Special:Linksearch – Find all external links to a particular site on en:Wikipedia; useful when a spam link is added by many different IP addresses or accounts.
    • External Link Cleanup – articles with excessive or inappropriate external links.
    • Template:Spamsearch - A template with lists of search terms that help to detect spam in the userspace.
  • For links that are generally used in an inappropriate way by new users, but which do not qualify for the meta or local spam blacklists, ask User:XLinkBot to monitor it with a request at User talk:XLinkBot/RevertList.