This post started as a reply to a post by Azuriel, but got out of hand quickly. So I made a blog post out of it. Since it’s a reply though, you won’t be able to understand much unless you go and read Azuriel’s post, which is a followup on Wilhelm’s post, which in turn looked at a report by SuperData Research. Go and read the two blog posts (they’re worth it) if you haven’t already done so, I’ll wait.

…

Now, I think Azuriel makes one mistake. The report mentions a data set of  36.9 million users, but it doesn’t say anywhere that all these users actually played a subscription MMO, let alone pay for any subscription in that time. I don’t think there is a direct relation between market share and those user numbers. What counts is the money. Look at the revenues, and you can try and “guess backwards” to figure out the number of users.

Now, I have to warn you: there is very little data available on exact subscription numbers and how they contribute to revenue. So at many points, I had to try and guess what I considered reasonable. It’s not quite as bad as a Fermi estimation, but the numbers have to be taken with a grain of salt.

WoW

An example: what is the average monthly spending for a WoW user? Last I checked, the US subscription rate was $15. Some players will pay more, some will pay less. That’s due to currency conversion rates, long-term subscriptions, or buying a sparkle pony. (The fine print seems to say that item shop revenue is included.)

If you now take WoW’s $1.041 billion, and divide that by $15, you end up with 5.78 million users on average over the year. That’s too low on users. Conversely, if you take the 7.6 million users that Azuriel mentions, you end up at an average of $11.41 per user per month. That’s too low on per-user revenue.

Or maybe it isn’t. There was probably a bunch of people on the Diablo III annual pass for a good chunk of the year, which they had paid for in 2012. I have no numbers on how many of those passes were sold, but I remember huge invite waves for the Mist of Pandaria beta (which was part of the annual pass package). So let’s just say 1 million users did not pay at all for their WoW subscription in 2013. That’s handwavey, because probably, most users who were interested into the annual pass probably bought it early on, and the annual pass was released in October 2012 (if I remember correctly). But that means even those people didn’t pay for subscriptions for 10 months during that year, so close enough.

That leaves 6.6 million paying customers. Which means $13.14 per user per month. Now we’re getting close. A 3-month subscription is $13.99 per month. Some will have 1-month subscriptions and pay more, some will have 6-month subscriptions and pay less. There will be some revenue from the item shop, but I can’t imagine that the big revenue generator for WoW. So this sounds reasonable.

edit: I was an idiot and off by a year on the annual pass thing. MoP was released in 2012, so handing out an annual pass in late 2012 with the prospect of a beta invite isn’t such a good deal if the expansion is already released… let’s look at the Chinese/Western split instead.

Or maybe it isn’t. WoW, more so than many other Western games, has a strong foothold in China. Or at least it used to have. The number thrown around is typically 50% of the players being in China. I tried to find numbers for that, but I couldn’t really, which is a bit disappointing. What we do know is that the Chinese market has fared considerably worse than the Western recently. There are two articles in Forbes from 2013 that Chinese players have been leaving in droves to other games. That’s good news for Blizzard, because the market isn’t nearly as profitable. Let’s just assume that, of the 7.6 million players, 5 million at this point are from the West and pay the monthly subscription. At an average price of $14/month, that provides a revenue of $840 million. Still $200 million short.

That leaves us with 2.6 million players from China. In China, players pay by the hour in the form of prepurchased game time cards. By poking around on the Chinese website, I found out that the often-quoted price of ¥0.45 per hour is still the current rate. Taking the average conversion rate for 2013, that translates to about $0.073. To produce a revenue of $200 million from 2.6 million players, they would therefore have to have played a total of 2.7 billion hours, or just shy of 3 hours per player per day. That lines up almost perfectly with the values from Nick Yee’s study. However, we have to be careful here, because those numbers are from a different point in time, on a different audience (Western vs. Chinese), and may suffer from selection bias. If I had to guess, I feel like 3 hours/day is a bit on the high side. Then again, there are revenue paths that I didn’t touch (server transfer fees, sparkle pony sales), which might make up for the difference.

EVE Online

A second example. Let’s look at EVE, whose user numbers looked way off by Azuriel’s estimation method. Their subscription is similarly priced, with longer subscriptions being cheaper than WoW’s. Then again, EVE gouges European customers more so than WoW does; in 2013, a Euro was valued, on average, at about $1.30; so EVE cost a whopping $19.50 per month on a 1-month recurring subscription, and still $14.30 at the yearly rate.

In addition, EVE’s subscriptions are supposedly funded indirectly for a significant part of the user base: older and richer characters with lots of income prefer to buy PLEX, which newer and poorer users buy sell for in-game money. However, PLEX comes at a premium: two month subscription worth of PLEX come at between $35 and $46 (=35€), so that’s between $17.50 and $23 per month.

Let’s make a really rough guess and assume that, on average, a direct subscription earns CCP $16 a month, while a PLEX subscription adds $20 to their coffers. For the next step, we’d need to know what fraction of accounts are paid by either option. I don’t think there’s any such information available. Sure, you can look at the amount of PLEX traded each day in Jita, but there are probably rich players playing the PLEX market, so not each sale will end up in 30 days of game time. Probably not by far. Besides, I don’t have historic trading volume values for 2013. However, these days, about 2500–3000 PLEX are traded in JITA each day, sometimes more. So that gives us something of an upper limit. (Station traders inflate the volume; however, back then it wasn’t unheard of to buy PLEX to fund your own subscription, because you could get in-game kickbacks from licensed PLEX-selling sites; those PLEX never showed up in the trading statistics.) Let’s say 60 000 accounts are paid by PLEX every month. That means about $15 million in revenue through PLEX.

That’s still not even close to the $93 million, of course. I was surprised myself how little of the market seems to be covered by PLEX. It certainly ties in with the claim of the “silent majority” that just flies in space for some missions and doesn’t have time or interest in making enough money to fund their game via PLEX. To cover the remaining $78 million, you’d need another 400 000 users by that calculation.

That gives us a total of 460 000 accounts, and that’s actually pretty close: in February 2013, CCP announced that they broke the 500 000-subscriber barrier. Who knows for how long they kept above the watermark; if the number of concurrent users is any sign, then probably not for long. (You can check those at Chribba’s EVE Offline.) But they probably didn’t plummet completely, either.