He says that:
"I keep saying the sexy job in the next ten years will be statisticians. People think I’m joking, but who would’ve guessed that computer engineers would’ve been the sexy job of the 1990s?"
So he's not a statistician then. Does anybody really think computer engineer was the 'sexy job of the 1990s'? Excluding Keanu Reeves in The Matrix? Still, we analysts can dream...
An ex-colleague once tried introducing himself as an econometrician to girls in bars. It didn't go well.
Back onto more serious matters, Varian also says that:
"On the Internet, we’ve learned to measure advertising effectiveness"
This is a common misconception and it doesn't seem to be going away. We can measure Web advertising effectiveness in the same way that Direct Mail has been able to measure effectiveness for ages. You send out a million direct mail envelopes (impressions) and get back ten thousand application forms (clicks.) Valuable, but not the whole picture.
Simple direct response measures like this say nothing about whether response rates are better when an above the line campaign is on air. They also can't help with whether the people responding are good prospects who would have responded anyway and you just gave them a convenient way to do it, or genuinely additional acquisitions.
These final two points are important and advertising agencies and Google (among others) are starting to recognise that web advertising response measures can't exist in a bubble. Web-based channels are going to have to get involved in statistical analyses of sales and borrow some of the techniques used to justify investment in TV. It's not just TV that can learn from Adwords, but vice-versa too.