Once again a WoW post. I'm sorry for the readers who aren't interested in it, but I find the question of why WoW has (had?) almost more players than all other MMORPGs combined too interesting to ignore. Eventually I'll blog about MMORPGs in general again; don't worry.
As you might be aware, there are two teams that worked on WoW over the years. Team A managed classic WoW and TBC and then moved on to Titan. Team B created WotLK and Cataclysm and is still in control. During classic and TBC the subscription numbers grew to over 10 mio. They abruptly stopped to grow in early WotLK when most players had finished raiding Naxx.
Therefore, being in Team B might not have been all that much fun during WotLK, anyway. But when sub numbers suddenly started to drop by 5% after Cataclysm was released, you can bet that Team B got nervous; as was whole Blizzard. Looking at the forums, the blue posters are irritable. Everybody there wants to know what went wrong; as does Trion or every other copy/paste company, by the way!
The question of why WoW suddenly stopped to grow, and now even declines, might actually be as interesting for the industry as the question of why it was as successful in the first place.
The WoW that was developed by Team A had specific ways of character advancement and loot distribution. Both were subject to severe changes by Team B.
The character advancement developed by Team A (copied from other games), included several raid tiers that required to be run through sequentially. First, players ran Molten Core, then Black Wing Lair, then An'Quirai, then Naxxramas. Several smaller dungeons and raids were available on the way.
The raid tiers were available to everybody as soon as they were developed. But if you started a new character you had to run through them in that order again - unless you wanted to be carried. Being carried, usually, was quite possible in 40 man raids.
The random loot system used by Team A allowed players to become well geared relatively soon. But getting the last final piece of gear could take for forever.
Team B, when they took over, were very enthusiastic. Who wouldn't have been? They took over the fastest growing money-printing machine in the industry's history. The genre was developing fast and many things seemed possible. In fact, the new neam planned to do a 7-chapter epic WotLK expansion.
In the end, not even half of that made it into the game. Did they just want too much? Did Activision later tell them that doing less would be more profitable? Who knows.
What we do know is that Team B set out to change the game to the better. Ghostcrawler started to communicate on the official forums extensively. Class revamps happened whenever feasible. The relentless focus on arena in TBC was scaled back. Most of all, raiding became more accessible.
For a long time players had complained that the classic:Naxxramas or later TBC:Sunwell raids were a waste of resources as only a few players would access them. Also, they had complained about the slow pace of completing a set. This was a (intended) consequence of the random loot system.
Most of all, there was the ever-present question, of how to make the endgame more similar to the leveling game. To understand the appeal of this question you must know that the leveling game works miraculously well. Even the WoW clones were easily able to provide good leveling games. But it also was quite expensive. Of course, Blizzard could easily have created a virtually endless leveling game with the one billion profit they made a year, but the management didn't want to reduce the return-on-investment down to the industry standard. So the designers tried to find a way to make the endgame more similar to the leveling game.
One way were daily quests. Another way was to increase the speed at which players got new items. The new raid-advancement mechanic would allow for that: Players would focus only on one raid at a time. They would chain-run dungeons to update their equipment accordingly. The Dungeon Finder would help them do this. By using a badge system, instead of random loot, players would not need as long to complete their sets.
Team B had only the best intentions. Everybody always has. And still they messed it up. Here's why:
Next to the phenomenal technical and artistic qualities, the central reason for WoW's success was that all players could always advance their characters by experiencing interesting content. That means that the content the players experienced was at a reasonable challenge level and happened in a pleasant social environment.
Since players are differently motivated, invest different amounts of time and maybe even are differently skilled, the game was flexible.
A player who played less often would experience the content at his own pace. If finishing a raid tier took him a year, then it took him a year. If it took him a month, it took him a month. All he needed to do was to find a compatible raid group or re-join one when he came back. Since raid groups were hard pressed to come up with 40 players and there always was somebody who couldn't make it that evening, random players would be invited. Part of the server community was built this way. It was how I got to raid the first few times.
The random loot system allowed players to get many items relatively fast. But even the players in the raid group that were most active, still had something to get from the lower-tier raids. In a way, random loot is evil and ingenious.
The change Team B brought along had a few advantages: Since everybody raided only one raid tier at a time, a large pool of players was available. This had been a constant problem for advanced raid groups in the old system. They often had problems finding replacements for doing advanced raids.
Also, the new system allowed players to replace items faster. Since this was more similar to the leveling game, it was considered a good thing (note: I disagree).
But the new system also had several disadvantages compared to the old one:
Since, at all times, all players raided the same raid tier, its difficulty had to be rather low. Otherwise, a lot of players would not have been willing (=able, if you insist) to do it. The new Naxxramas, consequently was rather trivial.
Since a lot of advanced players, who had superior network effects on the community, didn't like that, Team B had to create additional difficulty levels: heroic raids. These heroic raids, however, always felt highly gamey and artificial. Moreover, they were rather boring to complete, because players already knew most of the encounters. In contrast to what you read on forums, nobody plays MMORPGs for challenge alone. If someone just wanted a challenge, he played Mikado. The setting and the simulation aspect are important for MMORPGs.
Players who needed a year to complete content would suddenly not complete raid tiers ever. Players who completed raid tiers in a month would suddenly find themselves out of content.
Blizzard had always thrown away old content when they made an expansion, even Team A. Team B, however, even threw away raids in the same expansion: Since dungeons were much more efficient at providing better gear, players skipped older raids. The only saving grace were achievements. Most players, however, ran older raids only when they were seriously bored.
In combination with the Dungeon Finder and the daily dungeon quest, the dungeons became a grind in its purest form. While they did become more accessible, they also turned from a pleasant social experience into a boring, un-immersive, anonymous and at times asocial MUST. I'm sure it was easy to misinterpret the statistics. But trust me: nobody looked forward to do the daily dungeon. The pleasant social and reasonably accessible content that dungeons had once been was removed from the game.
The only reason to run dungeons was to later take part in raids. And since these raids were very accessible and relatively enjoyable in WotLK, sub numbers didn't drop. The rapid growth, however, stopped. Many players accepted the new dungeons and suffered through them to experience the story of World of Warcraft in the raids.
At the end of the WotLK expansion, Team B was mildly optimistic. Sure, their own expansion had stopped the rapid growth of the game, but they were about to change some things back to how they were. They would put back the challenge into the game! Unfortunately, they hadn't understood a thing.
Team A's WoW wasn't successful, because it was challenging. It was successful, because players advanced their characters while experiencing content at a reasonable challenge level and in a pleasant social environment. It was successful, because the game offered content for everybody to advance his character: from the most hardcore to the most casual.
Players were not herded together, but would play the game at their own pace.
To be able to make the game more challenging, Team B needed to make internal compromises. Being already in a defensive frame of mind, some developers sought support from the statistics department. There was a strong negative correlation between player deaths during leveling and the probability of re-subscription. Moreover, a vocal minority of players had loved to do the Loremaster achievement, which meant doing outleveled content. Nobody had the courage to argue against this. Consequently, the leveling game was made beyond trivial.
Maybe some developers still insisted that they needed an introductory raid, like MC, UBRS, Zul Gurub, Karazhan or Naxx. But being short on time and overstrained already, it was decided to skip it for now.
Heroic dungeons were designed to be very challenging early on so that they would still be challenging when players got better equipment. Unfortunately, this meant that shortly after Cataclysm's release the average player found himself confronted with a very trivial leveling game and very hardcore heroic dungeons that had often to be completed in an unpleasant social environment. And all this was necessary to start raiding which was rather hardcore and as such inaccessible for many of them.
The result was a 5% drop (after creative bookkeeping) in subscriber numbers shortly after Cataclysm had been released in Europe/NA and WotLK had been released in China.
To combat this drop, Team B, now 'supported' by several other teams, decided to get out more content faster. As a first measure the Zul'Gurub and Zul'Aman raids were modified and re-introduced as heroic dungeons. The player base complained, but overall this was mildly successful.
Moreover it was decided that an introductory raid was needed. The existing tier of raids was severely nerfed as the new tier was introduced. Unfortunately this didn't work out so well, because the then old raid tier competed with heroic dungeons. One lesson developers have learned by now is that players always prefer the easy and fast way to advance their character to the fun way. This usually also means that they quit soon after, complaining that the game is not fun. That's how players are.
In the end, there is only one thing developers need to do: They need to provide reasonably challenging content that players can experience in a pleasant social environment at their own pace. And they must do this for all players; from the most hardcore to the most casual.
Herding players together at one difficulty level by requiring them to regularly experience trivial monotonous content in an unpleasant social environment is the wrong way.
You're welcome to read the following posts on this topic where I switch from story telling to dry analysis. (1), (2).