Wikipedia hits logs

Brion Vibber (CTO of Wikimedia Foundation) just answered to one of my questions, with interesting links.

My concern was announcements such as this one, where it is reported that In August, 2007, Answers.com, a publicly traded company, announced to its shareholders that, due to a slight change in Google’s ranking algorithm, traffic and revenue for the company would decline by over 25%.
With the recent discussion over Knol, and rumors heard here and there, I was wondering if our hits from Google were possibly decreasing or not. If Google is really planning an online encyclopedia to enter in competition with Wikipedia, and want to favor theirs, they better twist the algorithms as early as possible if they do not want to be called « evil ». Well, even if they do not twist anything, such a project is likely to make our own ranks drop. As I already stated, this might have good side effects, with less financial pressure on us, and hopefully less jerks. But on the dark side, the press will raise hell, so, better anticipate 🙂

Anyway, Brion pointed out to one stats snapshot he made last summer. In short, 50% of our hits are coming from Wikipedia articles themselves, 25% from Google, 16% no referrer, 3% other wikimedia projects, 3% internal wikipedia search. All other referrers are 2% or less. This was from a sample taken last summer.

In two new graphics, he gave us the hits by referer group (%) and same, in raw sampled hit counts. This is from march 07 till february 08. Incidently, we are over 8 billions hits per month !

He reported that « the only significant change between March 2007 and now is a spike in May-June for ‘x-gadget:’ URLs which I believe are an IE 7 search widget thingy, which has steadily declined ever since. »

I also asked if there was any relevant way to follow up ranking of a word-search over time ? (as in « wikipedia result is in top 5 in january, in top 10 in march, in top 20 in june »).

One wikipedian (quite humorously) answered he had a brief discussion with A.L. about this many months ago where he suggested some kind of « Wikipedia World Domination Index », that is the percentage of terms that are right now in the public debate and are on 1st rank at google for a standard search.

Linguists from Leipzig are each day presenting the « Words of the day » based on the relative frequency of words in today’s newspapers.