I have a shocking theory about why Age of Conan and Warhammer Online had such a low retention rate, losing most of their subscribers after a short time: These games simply weren't very good. Shocking, I'm telling you! I'm arriving at that conclusion from economic theory, particularly the theory of the homo economicus, or basically the idea that people know what they are doing. Given the choice between similar games at similar cost, the majority of players will play the best game.
Of course that sends us down the slippery slope of what "best" is. Fans of smaller games often claim that World of Warcraft is the Big Mac to their games caviar. Unfortunately the comparison doesn't hold. People are not unable to tell the quality difference between a Big Mac and caviar, they are more influenced by the difference in price and convenience. A quick burger is often preferable to a nouvelle cuisine meal, in terms of price, quantity, and speed. MMORPGs aren't so different from each other, AoC, WAR, and WoW cost nearly the same, and have the same requirement of time. To assume that the larger number of players tends towards the *worse* game means you need to assume that the majority of players is too stupid to recognize quality. Which is a rather insulting and elitist assumption.
But how can we measure the quality of a MMORPG? Often the fans of one game point at features or game design the other game doesn't have. There is no denying that Age of Conan and Warhammer Online brought some good new ideas to the genre. But then they don't have all the features that World of Warcraft has, and comparing the length of feature lists doesn't work at all to compare quality. A better approach is to look at areas where the games are similar or even nearly identical. For example the auction house in WAR has a very similar functionality and game design as the WoW auction house. But if you look at the two side by side, it is obvious that the WoW auction house works much better than it's WAR equivalent. Or we could compare mob pathfinding, checking how likely a mob is to get stuck or be unable to find a path to the player. While we tend to focus on the differences, all these games belong to the same genre, and are by definition somewhat similar to each other, because otherwise they would fall out of the MMORPG genre. Thus if these games all have icons on hotbars to start spells and abilities in combat, we *can* compare the quality of execution of the different systems, check for example how reactive they are, or compare the visual quality of the icons.
Back in 2004 the fans of Everquest claimed that World of Warcraft won out against the nearly simultaneously released EQ2 due to better marketing. It is possible that there are undiscovered gems out there few people are playing. But that argument falls flat the moment a game gets a huge wave of initial subscribers, who *after* playing the new game for a while decide that it isn't for them. In that case either the gameplay is less appealing, or the quality of execution, the programming is inferior. Nobody would ever react with "Hey, this new game is more fun and runs better than WoW, lets go back to WoW". A customer who leaves and goes back to WoW means the new game failed to attract him. WoW might be the standard by which he measured that new game, but obviously he was willing to try something else, and would have staid if that something else had had sufficient quality.
This is important not because I like WoW better than WAR or AoC, that's a pure random event, albeit with high statistical probability. I could as well belong to the smaller groups of people who prefer the other games. But the important part comes when we move the discussion away from the Neanderthal-like "WoW bad, my game good, uga, uga!", and start to ask what exactly attracts some people to one game and other people to another game. Do these games have qualities that are diametrically opposed to each other? Or could we identify the strong points of several different games and create a game that is even better than WoW, by combining WoW's strengths with the strengths of other games. I recently mentioned that WoW for example is weak on social interaction, and some hardcore games are strong in that field, and proposed that a game could be both accessible *and* have strong social interaction if designed right. I would love to see some other games than WoW succeed better in the MMORPG market and get millions of subscribers. But for that to happen, people need to learn to analyze details. Nobody says WoW is perfect, but it must have rather good parts. The idea that WoW is universally bad, and millions of players have been too blind to see that after spending 4 years and thousands of hours, is downright ridiculous. That doesn't mean that only WoW clones can be a mass success, but it does mean that developers of future successful games do have to be willing to analyze what WoW did right, and which features of WoW aren't necessary for success. In the end a "WoW Killer" will come from a pool of ideas, many of which have been contributed by WoW itself, others from other games, and a few actually new ideas thrown into the mix. It would be foolish to dismiss ideas from any game out of stupid turf war considerations.