[teaspoon]

I don't know where this presumption comes from that Trends numbers are an accurate proxy for such-and-such metric. Even the more skeptical posts here seem to admit that Trends for "twitter" vs. "facebook" would mirror those services' market shares if it weren't for those net-illiterates googling for the Facebook login page.

[eli]

Well, I guess I'm even more skeptical then them because even removing people googling to get to a homepage, I think the numbers are only loosely correlated to traffic/popularity in any meaningful sense.