A reader asked me on my opinion on the announced guild advancement system in Cataclysm. By taking an inherently social entity, and adding rewards to it, guild advancement is potentially a very powerful tool of social engineering. Now some people have jumped to conclusions, and already dismissed it as clone of the suboptimal guild advancement systems that other games already have. But we simply don't know whether that is true yet, and one thing that Blizzard is famous for is stealing the badly executed ideas of others and unleashing their potential, thereby making their version much better than the original.
I recently read a book about management, which proclaimed as golden rule of managing people that "you get what you reward". I was laughing, because that sounded exactly like me talking about social engineering in game design. You get what you reward, and that does apply to guild advancement systems as well. Where previous systems frequently failed was that guild advancement systems more or less automatically rewarded pretty much everyone, as long as he was in a guild. That is still a possibility for the WoW system, but it would be a wasted chance. If three months after release both Ensidia and that pickup guild inviting random strangers both have the same level of guild advancement, then this system will simply fail to do anything. If your actions don't matter, and everybody gets rewarded the same way anyway, there is no effect of guild advancement on behavior. You get what you reward, and if you reward just being in a guild, you get people just being in a guild.
If done right, a guild advancement system can reward two desirable modes of behavior: Collaboration, and loyalty. But the way to there is strewn with pitfalls. For example it is far too easy to design a system where the guild advancement depends on a fixed number of points, and every action of a guild member adds to those points; with the inevitable result that larger guilds have an advantage over smaller guilds. You get what you reward, and in that case you'd get more large guilds, destroying many smaller ones. Another big pitfall is getting the encouragement of loyalty right: You want to reward people for working out difficulties with their guild, and sticking with their friends, instead of guild hopping. But you also don't want to punish them too harshly if ultimately things don't work out, somebody leaves, or gets kicked out, or a guild breaks up completely. Thus deciding what guild advantages a player loses when quitting a guild and joining a different one, and how fast he regains those advantages, is tricky. Too slow, and you lock people into guilds well beyond the point where the guild is still fun; too fast, and you end up encouraging guild hopping to the guild with the higher level of advancement.
So until I've seen the system in detail, and know all the parameters, I simply can't say whether the guild advancement system in Cataclysm will be good, neutral, or bad. Social engineering is a difficult science, and easy to get wrong.