[jim-greer]

I used to work at Pogo, EA's web games site. When I started it was all free, but we added a subscription tier with more features, more games, etc. When I left two years ago Pogo had about 15 million monthly uniques and 1.5 million paid subscribers (at $40/year, a great business). So our ratio was 10% - that was way above our expectations. Pogo has a very die-hard community - I'd plan on getting more like 5% if you execute it well.

If at all possible I would not charge for the existing free product, but instead think of new features that will be worth a premium to your top users.

[jacquesm]

That's exactly what we did at ww.com / camarades.com when the collapse of bubble 1.0 forced us to change strategies. Fortunately we had our plan ready to roll when the crunch came, we added a whole slew of (paid) features to the existing service, which remained free (and still is today). This allowed us to offset the drop in advertising revenue.