Monday, August 01, 2005
[Hat tip to SE Watch for the story, reminder sent in SE Report #105. :) ]
Claria's RelevancyRank, a search technology under development, challenges Google, Yahoo, and MSN with a comparison of where they would rank the sites that RelevancyRank (based on behavioral indicators that Claria gathers by following users around with its adware) would rank in the top ten for the phrase "cheap tickets."
The demonstration shows that MSN, Yahoo, and Google don't rank a lot of highly relevant sites (as far as Claria and its advertisers are concerned) in the top ten. For those scoring at home, Yahoo "won" on the sample query, by a score of 50 to 40 (Google) to 18 (MSN). Not too surprising given the strong business relationship between Yahoo and Claria!
Maybe the "failure" of search engines to rank behaviorally-compelling sites as high as Claria does isn't the point, though. On one hand, site quality can definitely be measured by some recipe of indicators based on fuller analytics, such as time spent on site, etc. On the other, certain metrics (such as commerce-related ones) might not be the best ranking criteria.
Probably, a fair comparison would also point to which paid ads are ranking high on the top three search destinations, also, since part of the point of organic search is to show users sites that are deemed to be "about something" and not just "about buying."
Still, it is a promising salvo by Claria, especially if most of the indicators are based simply on metrics like repeat visits, click rates, and time spent.
Back when Direct Hit came out, that was the sort of thing it intended to measure, and some of us thought it could vie for supremacy or at least wind up in the top four or so search engines. Unfortunately it had no good way of doing what it set out to do, and after being acquired by Ask Jeeves, its methodology got folded into the overall search mix.
Because all the search engines intend to deepen their use of behavioral analysis, Claria's take on search may not be as novel as it first appears. Google can follow user behavior in several ways: through the toolbar, through AdSense code, and through Google Conversion Tracker and Urchin installations. Finally, it now has personalized search with search history, and the idea of Google "accounts," meaning as long as you're logged into some element of the Google network, they can gather info about your search habits and what you click on for certain queries. If you wind up on sites that also contain Google code, or if you have the toolbar installed, they can gather considerable behavioral data, in theory.
Google, though, doesn't gear its whole algorithm to these behavioral indicators, as Claria proposes to do. This leaves a hole in search as far as Google's concerned. For users who would rather see sites that conform to Claria's idea of quality, there is no way to isolate on that element of the Google algo to allow those "behavioral quality indicators" to gain more weight. The one-algo-fits-all system (in spite of mild personalization initiatives to date) is potentially going to feel quite constraining to power users, and eventually, to mainstream searchers.
What that confirms for me is that a "bake your own algorithm" search engine will someday do very well, and this will be spurred on by various competing takes on search such as Claria's, Yahoo's My Web, and other upstarts who will potentially remind the public that there is more than one view of what counts as "relevancy." The idea that every so often, a Dance is done to create a new centralized edict about what constitutes the best formula for ranking pages on a given keyword query, is beginning to feel stale.
I think Google's days as the Pope of Search are numbered, if the innovations coming down the pike from elsewhere -- be they behaviorally-driven, peer-to-peer, or something else entirely -- are anything to go on. Big G will have to adapt.
View Posts by Category
Andrew's book, Winning Results With Google AdWords, (McGraw-Hill, 2nd ed.), is still helping tens of thousands of advertisers cut through the noise and set a solid course for campaign ROI.
And for a glowing review of the pioneering 1st ed. of the book, check out this review, by none other than Google's Matt Cutts.
Posts from 2002 to 2010