I'm sympathetic to some of Dave Roberts' critique of this year's Netroots Nation panels, but I'm not sure that this particular metric quite makes sense:
In some sense all conferences suffer from these same problems, but this one aspires to be a yearly event, so if it's going to continue -- and I heard they had trouble maxing out the 2,000 slots this year -- they need to get creative about offering something worth traveling and paying for.
It's of course flattering for an event to have it sell out quickly. But all selling out really proves is that you haven't priced your event correctly. A venue selling tickets to something wants to maximize revenues not maximize sales and the revenue-maximizing price is rarely going to be the same as the market-clearing price. I've got tickets, for example, to the sold out Rancid show on August 11 at the 9:30 Club. If the tickets had been 20 percent more expensive, I still would have bought one and so would a lot of other people. Maybe the venue wouldn't have sold out at that price, but they still would have made more money even if the place was left only 85 percent full.
Back to Netroots Nation, though, I wouldn't want to see them depart as radically from the bottom-up programming model that the conference currently works with. I think all they really need to do is maybe slightly reduce the number of panels that get approved and work a bit more aggressively to gin up submissions. Given adequate competition, it should be possible to put up a somewhat higher wheat/chaff ratio.