Lesson from "American Idol" Upset: Don't Put Too Much Stock in Google, Twitter

(MoneyWatch)  I've been spending part of the day obsessing about how Google, and Twitter, got the "American Idol" pick wrong, seeming to select, as of yesterday, the rather goth Adam Lambert over the fresh-faced Kris Allen. Given how social media and search are combining to create the National Focus Group, how could both of them be so wrong? Can the two of them be trusted to tell those in the media business, and other businesses, what the public wants? Should the people at NBC even care that the the Twit-ition to save "My Name Is Earl" (#saveearl) is currently the no. 1 trending topic on Twitter?

The answer on the Google question, according to 360i, the agency which published data yesterday on Mashable predicting Lambert would win, comes down to several factors:

1. Scandal. Apparently, photos of Lambert surfaced at some point in the last few months showing him cross-dressing and being friendly with other men, which may have caused his search traffic to spike. People who had no intention of voting for him, may, nonetheless, have wanted to go tabloid in their Google searches.
2. The difference between a vote and a search. 360i points out that searches are not votes, and shouldn't be viewed as such, especially when, as the agency's post says, a "13-year old girl can vote an unlimited number of times."

Interestingly, however, 360i points out that aside from Google, "votes" in social media are, in a sense, still picking Lambert to win even though the competition is over. As of this morning, he still had more than twice as many Facebook fans, and his first single is currently outselling Allen's on iTunes (though as this chart notes, three of Allen's songs are in the top 10, and a single from Fox's new show, "Glee," is outselling both).

All of which is to say, it's a bit too easy to get enamored with Internet data, and social media, partly because we didn't have such a visible, rich data mine before. As David Berkowitz, head of emerging media at 360i told me: "With any of this, there's a degree of reading too much into this." While I'm sure the people at NBC are watching the attempt to save "Earl" unfold on Twitter, they should keep that salt shaker handy -- they may need several grains of the white stuff. As much as Google got the "American Idol" pick wrong, there is even less known about how to project what goes on on Twitter out to the broader marketplace; it could be a bunch of "Earl" fans talking amongst themselves. Sometimes, Berkowitz, says, "All the buzz in the world is only going to do so much."

(UPDATE: The real-time search company OneRiot apparently predicted the "American Idol" winner correctly, using, as its metric, not volume but what people were saying about each contender. According to Mediapost, it sent a release out on Wednesday predicting Allen as the winner.