OT: By The Numbers

It is old news by this point, but I wanted to talk about the 600k subscription drop and what that actually means in the scheme of things. To be honest, my first reaction was “I told you so!” but without an actual breakdown of those 600k accounts no one outside Blizzard really knows what kind of people left – if the unsubs were from people who never zoned into a heroic dungeon let alone a raid, for example, then difficulty obviously would have nothing to do with that. Well… maybe they never zoned in because they were too difficult, but nevermind. In any case, here are some things to keep in mind:

Point 1: The numbers are actually significant.

The reactions among a lot of blogs and forum posters seemed to be almost dismissive of the numbers. “It’s only 5% of 12 million, /yawn.” While technically true, it is pretty inaccurate. Take a look at the following graph, which is a slightly modified graph you can see at MMOData:

The first thing you should notice, of course, is the huge dip that represents when WoW was banned in China for several months. It is worth noting because it indicated there are ~5 million WoW “subs” in China alone – the ~1.75 million subs still on the blue line represents the remainder of WoW East, which includes Korea and maybe Taiwan (the “does Taiwan count as China” deal is tricky business). With that in mind, here is Michael Morhaime:

Looking at the World of Warcraft side of the business, we were pleased to see record sales following the Cataclysm launch in the United States and Europe which helps drives growth and subscribership. During the first quarter of 2011, as players have eagerly consumed the new content, we have seen subscribership return to prelaunch levels in the West. We finished the quarter with more than 11.4 million subscribers worldwide. Moving forward, our objective is to continue delivering new content to players in all regions to further energize our community.

Key words: in the West. As in, that red line in the graph that has been largely stable since the release of Wrath. Now, there is nothing in the call itself that specifically says all 600k subs were solely from the West, but if they were, the drop suddenly goes from 5% to ~12% of anyone you or I could possibly be grouping with. Which leads me to my next point.

Point 2: WoW is not dying, but that is largely irrelevant to your individual experience.

I strongly believe people understand this point on a gut level, but sometimes get caught up in “logical” arguments over the internet. So picture this: did your daily WoW routine change at all when 5 million Chinese players suddenly could not log on? Assuming you are not Chinese, probably not. Ergo, anytime someone talks about 12 million 11.4 million WoW subs, they are really only talking about ~5.15 million WoW West subs that could possibly impact them in some way – using the bigger number just makes you feel better by identifying with a larger group, as opposed to it meaning anything in-game. It should even be broke down further into NA and EU, but that level of data is sadly no longer being kept by the MMOData people. A rough extrapolation from the chart would probably be ~3 million NA subs total.

And so that is the rub. If WoW lost, say, 2000 subs… but they were all from your server, that suddenly is a (personal) disaster. You either have to fork over some cash to transfer to other servers, or probably just quit the game. I do not know how many servers there are across NA and EU, but if we assume 400 total servers and 300k subs down (it’s possible the 600k drop came from post-peak Wrath launch in China), that still equals out to be 750 accounts per server. Drop in the bucket for Mal’Ganis with its 10k+ population, but a bigger deal on Auchindoun with our maybe 4k population.

Bottom line: any drop in your region is significant. WoW does not have to be dying overall for it to die for you.

Point 3: Ignore the vapid “there are always post-expansion peaks and drop-offs” argument.

It is literally true that more people buy the game/re-sub after an expansion is released than will be still playing the game several months later. However, I have seen this argument bandied about as if the exodus between Cataclysm’s release to post-Cataclysm was par for the course. Do I really need to remind people that Blizzard did not get from 0 to 12 million 11.4 million by having a 100% oscillation? At some point over WoW’s lifespan, it retained more players than it lost. Right here we have evidence that Cataclysm failed to retain as many players than it lost in the entire West region, e.g. where Cataclysm was released. Vanilla retained more players than it lost. TBC retained more players than it lost. Even Wrath retained more players than it lost. Cataclysm, thus far, has failed to do so.

That essentially sums up what I think about that.

A transcript of the Activision Blizzard investor call is available on Seeking Alpha for free, and I suggest reading it for yourself. Among other things, it has Michael Morhaime mentioning things we are starting to see now vis-a-vis even more “premium services” for WoW. For the record, I do not care that cross-realm RealID LFD groups can be formed for $3/month or whatever it ends up being. What I will say is that it is dangerously close to crossing into the Uncanny Valley of F2P-esque cash shop design where features that would have been free now suddenly cost extra money.

Calling it now, though: tri-spec and Dance Studio will cost $3/month. On the plus side, maybe that would get them to actually finish the latter. After all, we sure as hell get new $10 companion pets pretty regularly these days.

Posted on May 18, 2011, in WoW and tagged , , , , . Bookmark the permalink. 8 Comments.

  1. Is there evidence to support anything but the null hypothesis that there has been a flat 5% drop per server?

    Reading between the lines, I'm getting the sense you're implying that Chinese servers aren't losing players at the same rate, or even large servers for that matter, and that small servers are therefore getting hit disproportionately hard. It may be true, but is there any evidence of this?

    All else equal, small servers could suffer even in an underlying growth environment, due to network effects. Lose a couple of large guilds and you could have an exodus to other servers, in addition to abandoned accounts.


  2. “However, I have seen this argument bandied about as if the exodus between Cataclysm's release to post-Cataclysm was par for the course.”

    But, it is par for the course. Look at the chart above, focusing on the red line. It follows a roughly logarithmic growth pattern, but seems to hit an asymptote at around 5 million subscribers. Now look at the effect of BC and wotlk on that pattern. I don't see any effect whatsoever, the red line just continues on following it's logarithmic curve. By cata the line had leveled off at a constant 5 million, so the news that the current base is still 5 million is exactly what one would expect if Cata performed the same as wotlk and BC: no effect.

    You are arguing on the basis that BC and WOTLK retained more than they lost, which isn't a phrase I even really understand, but I don't see figures for retained or lost on that chart above, so you must be arguing based on data not linked in the article above. Until you link the data you're basing that statement on I'll have to let it slide, but based solely on the chart you posted and the recent Cata announcement I will indeed consider Cata “par for the course”


  3. If there had been losses in the low-value markets (China, etc.), Morhaime would have said that, to soften the blow.

    I conclude that ALL the losses were in high value markets (NA/EU), and that the losses in those markets may actually have exceeded the total losses, if there was an increase in China.


  4. @Midas

    Fundamentally, I cannot claim that all 600k sub losses came from just WoW West. It could have been split evenly, it could have been just 1000 subs lost in NA, it could be anything. What we DO know is that 100% of the sub gains in West from Cataclysm has been lost, which is signifigant regardless of the literal numbers. And remember, pre-Cataclysm levels means “14-months after ICC was released” Wrath levels, which cannot possibly be very good.

    @1st Anon

    It's worth noting that the red line and blue line are smooth growths while the combined green line is jumping all over the place. You will have to dig into MMOData for verification, but I'm going to argue that the red/blue lines in of themselves are meaningless extrapolations of the green subscriber numbers – they are smooth lines because MMOData knew there would be 8.5 million when TBC was release and 12 million when Wrath was released, with X% of that in West and Y% in East.

    When I say “retained more than lost” I mean exactly that: gaining 600k subs and then losing 300k is retaining more than you lost. Growth despite turnover, in other words. What is significant here is Cataclysm thus far has lost everyone it gained, achieving at a minimum 0% growth (and possibly 12% loss if all 600k were from West). I am not suggesting that growth only happens at the launch of expansions, but I find it hard to believe there is anything worse than anemic launch results.

    I guess we'll see how this plays out in the Q2 investor call.


  5. That's a very optimistic way to look at it and I think it's much worse. During prime time on my server there are rarely more than 10 people in the AH. I remember back in the days where you got disconnected in the AH because there were that many players…

    Do you have numbers on how many auctions are up on the AH now and since Cata launch? That should give some information on the server activity?

    That's why I assume that it's much worse:

    – Can the east actually lose subscribers? They pay per use and not per month. I would count every account ever created in the east as an active account. If they count like that we have lost more then 600'000 in the west.

    – Of course they also count all active account where player cancelled the renewal and don't log in anymore. I know that my account was still counted although I wasn't playing anymore. In the worst case the 600'000 was just from the first not-renewed accounts while still counting all active 6 month accounts. The number might grow during the next 6 month when more and more accounts expire.

    – They gave out 7 day of free play time to unsubscribed accounts last month. Would it be possible to count all those accounts as active accounts? Maybe that's why they did it last month, to fool the financial report.

    The problem is that this fueled a chain reaction. Everyone not logging in anymore is a person less for his friends to chat with. At some place his friends won't log in anymore because the chance of meeting someone online from your friend list shrinks and shrinks. Which further reduces the friends friends chance to meet someone online. And so on.


  6. The sky isn't falling.

    I could care less if we're back to pre-launch levels, personally. It actually makes game play better. In the couple months after Cata released, there were way too many players. And due to the way the design has continued to shoe-horn the population into small areas (capital cities and instances), this is even more relevant.

    Quite frankly, I haven't noticed a bit of difference aside from slightly less lag. And who really knows if that's due to the 4.1 patch (directx, network elements, etc), or if it's due to the population decrease. Regardless, my game play, and profit margins are still chugging along just fine. If anything, I have less competition, but the number of buyers seems about the same. Show me numbers and stats all day, but all I really care about is the observed impact.

    Games don't last forever. I have no doubt that Blizzard's timing for their next MMORPG (Titan) will be comfortably tied to the declining timeline of Warcraft. As a gamer, we still have more options now than ever before, and I believe that will only continue to swing in our favor into the future.


  7. Kring, my understanding is:

    Per-use customers in China are counted as active if they have played in the prior 30 days.

    Trial and seven day free accounts are not counted.

    It's my understanding that banned accounts still count for 30 days after banning; the theory was they do ban waves right before the end of the quarter (as they did in March) so some fraction of the botters will buy another account and inflate the numbers.


  8. That makes sense, thanks for the explanation neo.


%d bloggers like this: