John and I were just discussing the Clinton-Obama finishes in Iowa and New Hampshire, and how wrong the polls were about last night's primary results. Why were the polls and predictions wrong? What are polls worth if they miss the answer by that much? [Sadly, we didn't come up with definitive answers. Both of us are just amateur, if earnest, prognosticators.]

But the discussion got me thinking about focus groups and user testing vs. measuring actions. In the custom media business, we spend a great deal of time trying to figure out what readers need. Common sense says, just ask people what they want, and they'll tell you. In real life, it's rarely so simple.

Focus groups can help you get an idea of where your customers are. Or they could be derailed by a couple of folks with strong opinions. User testing -- which I'm a big believer in -- can give you a false picture of how people use your website. Or it might explain exactly why so many shopping carts are abandoned just before confirmation.

In the end, the only thing we can count on is what our customers actually do. Never mind what they tell us they want [more articles about the tax code! more lengthy educational pieces!] -- we must respond to what they actually read [more articles about Britney and Jamie Spears!].

Figuring out the most effective ways to measure and respond to customer actions, online and off, is our challenge.