I lean to the first bullet: Blame Comscore. As some are pointing out, and as Comscore's aftermarket weakness is showing, this is turning into an acid test for Comscore -- and it's failing.
The funny thing to me is that ANYONE ever believes ComScore any more. Not only have they been demonstrably wrong on query counts, search share, and click growth before, but the people who run Google, Yahoo and other publishers have said as much.
Paid Clicks - Aggregate paid clicks, which include clicks related to ads served on Google sites and the sites of our AdSense partners, increased approximately 20% over the first quarter of 2007 and approximately 4% over the fourth quarter of 2007.
Here's Anand from datawocky describing comScore's methods.
ComScore sets a lot of store on their "panel-based" approach, which collects data from a panel of users, similar to Nielsen's method of collecting data on TV viewing using data from a few households that have their set-top boxes installed. ComScore has been in this business longer than anyone else, and has arguably the best methodology (i.e., algorithm) in town to analyze the data.
I don't think they have the best model, simply because it's been wrong many times before. Here are a few of the ones I've documented in my blog - not a comprehensive list of all of their mistakes, since I gave up whining about this in 2007...
You can find other bloggers who discuss comScore's data in 2007. I believe that comScore has never been very good at analyzing the traffic or click patterns of top 10 web properties like Google, Yahoo and MySpace.
comScore did release a press release two days before the earnings, which seems to provide some cover contra Kedrosky's off-the-cuff "blame comScore":
The data from Comscore, which was not released publicly but provided by Wall Street brokerages, was seen as welcome news for Google after a difficult month in which many financial analysts lowered their expectations for the search company's first quarter.
Comscore last month prompted concern among investors and analysts when it reported weakness in the number of consumers clicking on the Internet giant's search ads in February.
The explanation that the data "was provided by Wall Street brokerages" doesn't make much sense as written - probably comScore secretly provided a CYA set of data TO Wall St... who knows?
Just remember this the next time comScore reports on some key metric.