My last two posts on Wikipedia Wikipedia is now the Web’s third-most-popular news source….For Woman? and The 5 W’s of Wikipedia - Who, What, When, Where and Why of the search for Wikipeida. part 2 covered who was searching for Wikipedia but what about Search Clickstream Analysis - the stuff of goes with Visitor Behavior and visual overlays that is the domain of Web Traffic Analysis?
Above, I took the top terms searched for before getting to Wikipedia (left) and then those terms searched for as a result of looking for Wikipedia on the right. Again, we’d need the long tail of all the searches to and from to get a complete picture.
First thing, there’s less searches that happen after looking for Wikipedia meaning that ~9,400 searchers got what they needed and did not search any more (about 34% if we include the 2.54% looking for Wikipedia.org). The rest (those that still felt they needed more information, a surprizing amount go to Ask Jeeves! Actually, it should be Ask.com but people are still using the old name for the search engine.
This chart above looks much like HitWise’s clickstream of a site using search terms, but they have more precise data to work with (and you have to spend a hell of a lot more money to get that data). All the data I pull is free - it’s the creativity and art that’s added and.
So if you’ve got the money, go with HitWise - but if you don’t or can’t, use many of the free tools that are available and when you add intelligence and creativity to the data, you can get results that sometimes rival or surpass what HitWise produces.
It appears, based on the Search Volume Seasonality Forecast that search demand for Wikipedia is going up and up and up. But since the data that Microsoft provides is only though last year - it’s not helping us much now.
Maybe Google Trends can be of more help.
 |
Wikipedia movement meets in Frankfurt Expatica - Aug 5 2005 |
 |
Science Journal: Wikipedia Pretty Accurate phillyBurbs.com - Dec 15 2005 |
 |
Court Orders Part of Wikipedia Off-Line phillyBurbs.com - Jan 20 2006 |
 |
Wikipedia Surpasses One Million Entries Digital Divide Network - Mar 2 2006 |
 |
Group adds to Nussle’s entry in Wikipedia DesMoinesRegister.com - Mar 24 2006 |
 |
Wikipedia Ripe for Political Dirty Tricks CBS News - Apr 28 2006 |
And the news stories that drove traffic are also listed. Overall, more demand for Wikipedia came from Germany and Finland than the US (strange).
Looking at Alexaholic.com we can see the traffic demand of Wikipedia as related to some of the other sites that people also use (ask, myspace, amazon, etc).
Based on Pageviews, MySpace outpaces everyone else. We don’t know how accurate Alexa’s data is because it’s skewed and that’s been covered in a couple of Matt Cutt’s post like this one.
Finally, lets look at some statistics that I published earlier on Wikipedia
………. Even for Wikipedia (the gold standard of the genre) half of all edits are made by just 2.5% of all users. And note that in this context user means “logged in user”, not accounting for the millions of lurkers directed to Wikipedia via search engine traffic for instance.
In fact, Wikipdedia follows the 80/10 rule where 10% of all users make 80% of the edits to Wikipedia, but of that 10%, 5% make 66% of the edits and half of the edits are made by 2.5% of all the Wikipedia users. Anonymous ip addresses make edit about 18% of the Wikipedia content with most of the editing done on Articles (85%) and Talk Pages (8%) based on 2003 and 2004 numbers.
The authors of a report to the US Congress on Wikipedia last year also dabble in user base Demographics (similar to the Mosiac system that’s now part of MSN AdCenter) and say that…
- Social types - Socialites ant Trolls
- Article types - Worker Bees, POV pushers
- Policy types - Police, Judges
- Controversy lovers - Moths
- Pseudo-users - Sock puppets and Vandals
- Extra-Wiki - Mailing list, IRC, Board activities and Developers
I think in these last three posts (including this one) I answered the 5 W’s of search for Wikipedia, as best as can be answered (without going into another round of posts).
