27

I understand that rel="nofollow" is a safe way to link to external sites but why would Wikipedia refuse to link to other sites without the nofollow?

Is this a matter of trust with the community?

Wouldn't there be something good in SEO for Wikipedia if they used follow links for trusted sources?

See example below: enter image description here

53

One of the ways search engines determine what Google calls “PageRank” is by the number of incoming links, with more weight given to sites with high page-ranks (NB: I will use this term as the generic word to mean how a search engine prioritises how the results are displayed).

This has resulted in so-called “link spam” where multiple sites have been created for the purpose of linking from them to try and boost the page-rank of the linked site.

Since Wikipedia is one of the most highly ranked sites on the internet, and it allows anyone to create links from there, it would be a ideal place for people to create pages and links for the sole purpose of boosting their site’s page-rank.

By enforcing “nofollow” tags, it tells the search engines that these links are worthless when it comes to considering their value to the linked site. This means the incentive to fill Wikipedia with link-spam is basically eliminated.

  • 12
    Correct. On the other hand, it also gives an excuse for Wikipedia not to pass on ranking to their sources which are arguably important to the credibility of Wikipedia. – Itai Jan 17 '18 at 18:55
  • 8
    Not sure “excuse” is the right word here. I don’t think WP is deliberately trying to stiff their sources. I believe that it is a consequence of the solution to one problem creating a new problem. Unless there is a way to prevent link-spam in source links, I think it’s just part of the compromise we have to make to mitigate the prevalence of link spam site-wide. – Dwev Jan 17 '18 at 19:03
  • 26
    Maybe I’m missing something. It’s serving their millions of daily visitors to not provide bad pages full of link spam. As a user, I’m grateful for that. If link spam had to be manually moderated by editors, they wouldn’t have time to do anything else, and WP would just become another link farm... what am I missing? – Dwev Jan 17 '18 at 19:10
  • 5
    They could remove nofollow for links that have stayed on WP for a significant time - if they weren’t removed they’re likely not spam. SO and Reddit do this using upvotes as a measure. – DisgruntledGoat Jan 18 '18 at 10:49
  • 3
    @DisgruntledGoat While that seems like a good solution, I'd expect blackhat SEO vendors to quickly start buying commonly-linked domains or paying their owners. – ceejayoz Jan 18 '18 at 13:50
4

There's a couple of reasons why any website (including Wikipedia) that welcomes user content today would make links as no-follow.

  1. Links can become outdated and may even return error pages (example: error 404) to the user when accessed. A web page that contains at least one link to an error page is not a good page to be indexed on the Internet.

  2. As someone mentioned, websites that allow users to publish content can potentially manipulate the ranking of the site and/or other sites by advertising links to any website (depending on what filters the website the guest is using has in place). As a concrete example, If a user publishes hundreds of links to vulgar and/or sex when the site has no relation to that, it may at the very least confuse search engine robots into believing the website has more to hide and that it can't be trusted on to deliver quality results that users expect to see.

In order for webmasters to protect themselves from fooling search engine robots, they will (or at least should) do one of the following when having user content published for the public to see:

  1. They will put the content on hold while an actual person checks the content to see if its acceptable for the site then publishes it manually. This is the best option, however we are in the world where people don't have time to check thousands of paragraphs of content.

  2. They will let the content be published right away but make every single link as no-follow, at least until the links are manually validated.

  3. They will use automated filters to see if the content matches against certain words or phrases and if it does (or does not), then content will be rejected.

In any event, if the link doesn't add any value to any user or is not related to the website in any way, it should be made as no-follow, or better yet, the link should not be available at all.

  • 6
    "1. Links can become outdated..." - this can happen to any link, anywhere. This isn't a reason to use nofollow. – DocRoot Jan 18 '18 at 9:59
  • @DocRoot if a search engine can see that a website owner isn't maintaining their site to deal with link rot or otherwise has invalid links, they can dock that site points in the ranking. (It's a poor user experience if they keep getting linked out to 404s; Google et al. want to be trusted to guide users to websites where they are likely to have better user experience.) Without nofollow, someone could take Wikipedia down in the rankings by spamming it with a lot of invalid links. – WBT Jan 18 '18 at 17:28
  • @WBT Yes, but that isn't a reason to make the link nofollow to begin with - in anticipation of "link rot". There are (automated) ways to validate the link target for 404s. I'm not saying the Wikipedia links shouldn't be nofollow (for the sake of spam prevention), but not for point #1 as mentioned above. – DocRoot Jan 18 '18 at 17:38
  • @DocRoot How about in anticipation that someone might try to reduce your search engine ranking (perhaps to favor their own site), by posting links that are invalid to begin with, not just rotting over time? (Or, which are selected to intentionally rot out after a few minutes, if there's some initial validity check?) I think Wikipedia's playing defense here. – WBT Jan 18 '18 at 17:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.