The comments in this article baffle me to a point where I question whether it was stated at all. Gabe Newell was quoted recently saying,

"When we talk to partners who do free-to-play, a lot of people see about a two to three per cent conversion rate of the people in their audience who actually buy something,” said Newell speaking at the North of Innovation conference in Seattle, as quoted by Geekwire.

“Then with Team Fortress 2, which looks more like Arkham Asylum in terms of the user profile and the content, we see about a 20 to 30 per cent conversion rate of people who are playing those games who buy something.”

Newell added: “There seems to be something about the content that significantly changes how your monetisation occurs, with apparently much broader participation than you would see out of something like FarmVille.”

The Valve boss said his studio doesn’t fully understand why this is, but it will continue to run experiments to better understand the customer data.

Why is there confusion here? Valve took a core-audience game they charged around $20 for and made it $0. Why is it surprising that a player would spend more money on a great game like Team Fortress 2, one that was once valued at $20 when compared to a basic title like Farmville which is and always was free? I agree that 20-30% penetration is pretty impressive but any increase makes sense.

He states that the "content" makes a difference as some sort of revelation. Somehow they're surprised that if the add-on content is crap, no one will buy it? In Valve's defense, the article does state that they don't, "fully" understand the trend.

Stumbled upon this over at Develop.