I have long pondered the notion of self-funding staff via microcontent. For example, hire more graphics artists for City of Heroes costume designs and then sell their output for a few dollars as an account upgrade. If they sell enough, they stay on. If the model works really well, you can open it up to independent contractors who could make and sell costume pieces, and otherwise replicate a bit of what goes on in Second Life with user-generated content. With the host company taking a percentage. See Julian on fluff for other thoughts.
Yes, I know there are issues there. I liked the idea of putting together Dungeons and Dragons Online dungeons that way, but you would very quickly run into balance issues where dungeons were intentionally too easy or rewarding, because people would buy them for easy xp/loot. See power creep in all the existing D&D books. (After all, why buy the new book if it does not have a prestige class you want to use, and why use one that is weaker than existing ones?) (Yes, I know.)
City of Heroes has kind of pulled this off. Lead developer Positron comments:
Well, the Villain Epic ATs were originally planned for I13, but the brisk sales of the Wedding Pack enabled us to fast track these by getting them budgeted to be done earlier. Every time you see a new tux or wedding dress you can send a thanks to that player for getting everyone VEATs an issue early.
I may be misinterpreting, but it sounds like CoX bought itself more staff time via microcontent. That is the opposite order, but it works as a potential model. Most of my CoX friends bought the microcontent for the jump pack and the Pocket D teleporter.