On Fri, May 21, 2010 at 2:06 AM, Andrew Gray andrew.gray@dunelm.org.uk wrote:
We have a rough proxy for use of search, I believe, by seeing the traffic to special:search. We could try seeing if this exhibits any noticeable difference after the chance.
We're limited here by the fact that, for privacy and other reasons, we don't track clickthroughs (so we can't tell how many people are arriving at an article from the search box, only to go back to searching).
That said, according to stats.grok.se, pageviews for Special:Search dropped below 5 million hits per day three times in the last week, having not done so since January 3 this year.
On the other hand, you could interpret any such change in about seven different ways, so it may not help much... ;-)
Indeed, that's why it will be interesting to hear what metrics the interface people used when making the decision to carry out this change, presuming that they did actually use some.