Comments on: Wikipedia and Google

By: How The Associated Press will try to rival Wikipedia in search results » Nieman Journalism Lab

Thu, 13 Aug 2009 12:38:05 +0000

[…] results, which is a product of all the external links pointing to Wikipedia and a variety of other factors. As Mathieu O’Neil, the editor of in-house newsletter Wikipedia Signpost, told me yesterday, […]

By: Jimmy Wales

Jimmy Wales — Sat, 11 Jul 2009 17:01:31 +0000

Thank you Mathieu, and yes, the way that quote is normally attributed to me is exactly the opposite of what I believe.

By: Mathieu O'Neil

Mathieu O'Neil — Fri, 10 Jul 2009 11:57:32 +0000

OK… you seem to be using the example of a famous racehorse as an illustration of the fact that Wikipedia pages with lost of internal links (including links to pages such as “California”) will garner index authority… well, maybe. But you ignore one all-important fact here, I think. We are dealing with the ponies, racing, bets. Ring a bell? That’s right – it’s obvious the Camorra, the Russian Mafiya and the Hong Kong Triads are lubricating the web with Google Juice in order to, ah, manipulate something or other. Seriously though I did think that PageRank was supposed to measure links between sites not pages… will have to check I guess.

By: Sage Ross

Sage Ross — Fri, 10 Jul 2009 04:26:09 +0000

PageRank, at least in its basic concept, pays no mind to whether a link is coming from within the same website or from an unaffiliated website. It deals with the individual webpage as the unit of analysis, so each article on Wikipedia has its own PageRank. Because most of the links from any given article are to other Wikipedia pages (nearly all since Wikipedia went “nofollow”, which happened well after it rose to search engine dominance), Wikipedia is very efficient at distributing ‘Google juice’ amongst its articles. But that comes ultimately from the very high number of links coming to Wikipedia from the outside; in large part, Wikipedia’s search engine dominance can be originally traced to bloggers. (It’s not just Google, either. Wikipedia has similar prominence in the results of other search engines.)

Much of search engine optimization consists of more deliberate efforts at apportioning the ‘Google juice’ for a set of webpages (whether in a single site, or across multiple sites controlled by the same entity). For example, SEOs will deliberately make sure many pages link to those specific pages they want to boost the PageRank of while at the same time limiting the number of other links so the influence of the remaining links is more concentrated.

Wikipedia, in contrast, spreads its links out and has a “natural” (if that word can apply here) link structure that takes all the incoming ‘Google juice’ that accrues to externally popular articles (the ones bloggers and others link to) and reapportions it according to internal popularity (the articles that other articles link to).

You can see cases where PageRank and Wikipedia’s internal linking structure complement each other to result in atypically low search results for bad articles or ones that are tucked away in isolated corners of Wikipedia. For example, do a Google search for “arts and letters”, go through the results until you find the Wikipedia, and then take a look the article and find out why.