To some extent virtual items sales were scorned as a result of a very player desire for what you could call 'fairness'. Very many of the people who made up the MMORPG playerbase at the time would have been strongly opposed to any sort of virtual good transactions, and even now there is strong opposition to it when developers consider adding it to their games.
For example, CCP (developers of Eve Online) have recently begun experimenting with microtransactions, and were met with a strong outcry from their playerbase at the first sign of any purchase that could remotely affect gameplay. The outcry was strong enough that they backpedaled completely, as you can see here: http://www.eveonline.com/devblog.asp?a=blog&bid=815
Now, this example is (I think) interesting in two ways:
First, by listening to the player outcry, CCP is blocking themselves off from a very large potential revenue stream. Traditionally EVE Online has a tough time acquiring new players because it is famous for its extremely high learning curve and brutal 'fairness' in which players are subject to non-consensual Player versus Player combat in almost any part of the game world. Selling in-game advantages for money would potentially let them tap into a market of people who like the game but don't like being at a disadvantage compared to more experienced players. One could argue that Zynga's success was not due to believing in virtual goods, but due to their success at finding a new market that happened to have no opposition to virtual goods.
Second, the player outcry demonstrates that while the success of companies like Zynga has helped make it possible for game developers to turn a profit using virtual items sales, there's still a large number of players out there who can't stand the idea of selling in-game advantages and even resent the sale of cosmetic items. As a result, switching from subscriptions to microtransactions is a risky prospect for most established online game developers, and the bigger examples (like Turbine's Lord of the Rings Online going free-to-play) are cases where a developer moved to microtransactions out of desperation in response to declining subscription numbers. Even if they can make a microtransaction model work, if they alienate too many of their existing customers, they may not be able to stay solvent long enough to follow through on the transition.