Josh Kopelman says the first penny is the hardest.
The truth is, scaling from $5 to $50 million is not the toughest part of a new venture - it’s getting your users to pay you anything at all. The biggest gap in any venture is that between a service that is free and one that costs a penny. (emphasis in original)
Jeff says "people will pay for these services but they have to have some persistent utility in order to cross that threshold, but the persistent utility is what is challenging."
Converting free users into paying users is only challenging if you don’t ask them to pay. Premium feature sets is one way to do this, PBWiki being a good example, while increasing consumption thresholds being another. DabbleDB gives you more users for each pricing tier, other companies give you the ability to create more files, etc. Basically, the latter is what I prefer because it gives new users the opportunity to appreciate all the features while at the same time putting a limit on how much they can consume.
Josh argues against the the "premium" model for this reason, and instead looking for other ways to monetize the audience.
At some point, the cost to acquire a paying customer is so high, it makes sense to consider shifting from a pay model to a free model. In these cases, asking "who would pay to reach these consumers" (or "who can subsidize these users") creates an opportunity to build a more valuable business through the combination of exponential growth and targeted advertising.
"Premium" services can certainly be a sustainable business model, but the ones that truly change the game are the ones can harness that user base (and preferably in a more efficient manner than just advertising). Those are the ones that, as Josh puts it, shrink the market. I haven't discussed what we're working on publicly, but that's precisely what we're trying to do - I can't wait to talk about it more in the coming months.