Blog Archives

Inevitability of Decline

Nils and I have been debating here lately over whether the decline of WoW’s growth had something to do with Wrath and Cataclysm’s design. One of my biggest pet peeves when it comes to WoW discussion is the notion that subscription numbers are somehow correlated with endgame design. They are not. As I mentioned in that article, only ~20% of subscribers killed the easiest (non-Naxx) boss in the entirety of Wrath and the stats for Cataclysm thus far¹ are not any better: 17.9% have killed Magmaw, even after the nerf. But my actual argument goes further than that:

The decline of WoW was/is inevitable. That is to say, WoW’s subscription growth would have slowed and eventually declined no matter what Blizzard did. This is in contrast to the implied argument from Nils and others that had Blizzard simply copy & pasted TBC, they would have gained 2 million subs per year into perpetuity. My argument about the inevitability of decline has further consequences when it comes to MMO game design, because I believe that good design decisions can still lead to net loss in subscribers. Bad design decisions can certainly increase the magnitude of the hemorrhaging, but the best you can hope for with good design is to stave off the inevitable as long as necessary. This argument rests on a few premises.

Premise 1: Fun has diminishing returns.

I explored this premise a bit in The Diminishing Returns of Fun post. The basic idea is that Novelty is the ineffable quality of “newness” of a game that is consumed as the game is experienced. It can also be expressed by the quotes “You can’t go home again” and “You can’t step into the same river twice.” Novelty can be about learning new systems (i.e. rules) within a game, but it can also be about seeing new areas². Part of the “ineffable” problem is that novelty is not just about new things in of themselves. A jigsaw puzzle you previously solved does not become novel simply because new rules are introduced, like forcing yourself to not start with edge pieces or playing Tic-Tac-Toe before being able to place a piece. Once you feel confident of the contours of an experience, the remaining novelty quickly evaporates.

Nils described this once as: “The longer you play a game, the lower its potential to keep your mind busy, because you get ever better at it. If you so want, the game is in the cache now and you don’t have to think as much to play it.”

How this relates to WoW can be expressed in this investor call quote from Blizzard president Michale Morhaime:

“As our players have become more experienced playing World of Warcraft over many years, they have become much better and much faster at consuming content,” he said at the time. “And so I think with Cataclysm they were able to consume the content faster than with previous expansions, but that’s why we’re working on developing more content.”

Players become much better and faster at consuming content because the novelty of said content has already been experienced. A quest that might once have been novel (despite being built from un-novel components like Kill X Foozles, etc), becomes much less so the Nth time around. With the novelty gone, there is no compunction against finishing the content as quickly as possible – the quest becomes a task to be completed, instead of an experience to be sensed. Anyone can open a book to its final chapter and read how it ends, find a plot summary or otherwise “spoil” it. That we do not do this is a function of our desire to experience the story, which (typically) only occurs by limiting ourselves to reading it in order from beginning to end.

Unfortunately for Morhaime, developing more content will not stop the decline, no matter how good it is.

Premise 2: Market Saturation exists.

As I talked about in Saturation, Tom Chilton made the intriguing comment in an interview that:

” […] if you look at the way the population breaks down, we’re at a point in our history where there are more people that played World of Warcraft but no longer play World of Warcraft than currently play World of Warcraft.”

If measured at the peak of WoW’s reported subscription rate of 12 million, that means roughly 24 million people have played WoW at some point in time. How many more people would play WoW that have not done so already? That is somewhat of an open question. However, if we look at the Wikipedia list of best-selling games of all time we see that more people have bought World of Warcraft than Halo 1-2-3 combined (5m, 8m, 8.1m), more than Super Mario World (20m), more than Super Mario 64 and Mario Kart 64 combined (11.62m, 9m), more than The Sims (16m), and damn near two-thirds of the way of overcoming the original Super Mario Bros that came with every single NES (40m).

I doubt that Chilton was including Free Trial downloads in his statement, but either way, it is difficult to believe that the market has been anything other than tapped. If we assume that Blizzard is staffed with rational bussinessmen (if not designers), then we can infer from Chilton’s statements that the market for WoW has peaked according to Blizzard’s own data, and further sub growth is more likely to come from additional localizations than, say, capturing/retaining more US/EU subs.

Premise 3: Players consume content faster than designers create it.

The difference between this and Premise 1, is that Premise 1 is about how it becomes increasingly difficult to create subsequent content as novel as the original. This particular premise is simply how long it takes to create content versus how long it takes to consume it, novel experience or no.

Factoid: There were 8.75 million subscribers in vanilla.

Back in my Subscription and Correlation post, I augmented a graph from to show the release date information of the WoW expansions for illustration purposes. This is what it looked like:

The interesting thing to note is that the total subscription numbers (the green line) is actually above 8 million before the release of TBC. China did not get TBC until the end of 2007, so if you add that interim period (the blue line) to the whole, we get WoW’s vanilla population at 8.75 million.

I believe this factoid is important because that 8.75 million segment can be considered the Baby Boomers of WoW. Indeed, between 2006 and 2007 the sub population grew from 6 million to 8 million – churn notwithstanding, it is possible that 6 million subs aged for an entire year. Compare that with 2005 to 2006, where less than a million subs could be said to be one year old by the end. In other words, starting in 2006 the volume of veteran players more than likely outpaced new players.

Premise 1 + 2 + 3 + Factoid = The decline of WoW was/is inevitable.

Ultimately, I wanted to stake out this argument because A) I do not see many people that do despite implicitly acknowledging it (e.g. growth is always finite), and B) good MMO design decisions tend to be indistinguishable from the bad when viewed through the one-dimensional prism of subscription numbers.

For example, the general (blogging) zeitgeist surrounding Wrath of the Lich King was that it killed the goose that laid the golden eggs. What does not get separated was, say, how the questing experience was orders of magnitude better than questing in TBC (does anyone seriously disagree?), and how I feel that the episodic raiding model is a lot more honest, and better designed than a strict linear progression model. Yes, raid difficulty was poorly handled. Yes, turning heroic dungeons into 5-man daily quests burns people out (starting with TBC, mind you). None of those raiding missteps can realistically can be responsible for more than a fraction of sub losses, but that is neither here nor there.

The mechanics of what I am asserting can be visualized in the crude image below:

In frame 1, a player has just begun playing. In frame 2, they have carved a path through the majority of the game on their way to the level cap. In frame 3, they have reached the endgame, where most of their activities involve repeated content such as dungeon runs, raiding, dailies, and so on. In frame 4, the player either rolls alts or otherwise backtracks in search of novel experience in the content they missed (different leveling zones, class-specific quests, the opposing faction, etc). As expansions are added, the frames get taller… but not by much. Whereas the original leveling experience might have taken 300 hours, leveling in an expansion takes a fraction of that. If an expansion is released when you are in frame 2 or 3, that is fine. But if you are in frame 4… the end is (already) near.

So, in summation, WoW would have inevitably declined no matter what Blizzard did. That the decline “began” during Wrath is largely an irrelevant coincidence compared to the Baby Boomer population wave reaching the natural end of their novel experiences. The Baby Boomer hypothesis can be falsified should we ever get average age of account statistics or character maps of activity, but it would not affect the soundness of the underlying argument either way. Raid design decisions are unlikely to have anything but marginal effects on subscriptions compared with what the bulk of players are doing – which we know to be not raiding. And I believe that a lot of better game design could be achieved if we spent less time fixating on a drawn-out, endgame experience.

The deserved popularity of WoW proper came from the strength of its IP narrative, its pacing, its humor, the vastness of its game world, the underlying character of each of its zones, the uniqueness of its classes and how each demonstratively created their own novel experience. We should not take its declining subscription rates as anything more than the natural decline of an otherwise well-lived life. Here is to hoping that Blizzard opts for hospice care instead of the intensive life support that is currently in vogue.

If not… well, here is a raised glass to Diablo 3 and Titan.


¹ WoWProgress says 64,642 guilds have killed Magmaw as of the time of this writing. Assuming 18 raiders per guild (charitable considering 10-man raiding is vastly more popular), that means 1,163,556 players. WoWProgress only tracks NA/EU/KR/TW guilds, which number ~6.5 million. Ergo, 17.9%.

² You can probably argue that exploration equates to “learning the rules.”