Three headlines today: Runescape Hits 100,000 Subscribers; Final Fantasy Grows Enormously; and Girardo Develops Cool Tracking Stats...
From the "Another Win For Innovation Over Graphics" Department: Runescape runs on Java, and now has over 100,000 subscribers. Although gameplay doesn't look dramatically different from anything we've seen before, the business model is unique. You can play for free if you want; pay $5 monthly to get access to more quests, skills, storage space. And it's from England. Auf geht's, Europa!
Meanwhile, Bruce Sterling Woodcock's latest MMOG growth chart shows Final Fantasy XI piercing the 500,000-subscriber barrier, on a trajectory whose second derivative hasn't gone negative yet. Are we about to see the first truly million-person virtual world? (Lineage's huge numbers are tainted by PC-Baang subscriptions.) By the way, Bruce, if you're listening: we are all in your debt for these numbers, many thanks.
And finally, have a look at Luca Girardo's new MMOG site. Coolest thing: a chart showing web traffic to the homepages of major virtual worlds over time. The numbers are newly available because Amazon.com is trying to out-google Google, and unlike the soon-to-be-IPO'd giant, they freely show the site statistics that generate their rankings. Here, for example, is Terra Nova's datapage. Not sure why we grew in March and shrunk in April, but we did. What's cool about the Girardo Index is that it measures buzz, even for worlds that are only in beta. One finding from the current figures: it looks like World of Warcraft is going to be a fairly big deal.
I'm not sure why Lineage's numbers are tainted by PC Baang play. Does it really matter to us as observers how much people are paying them or is it more important to judge how many people are playing them? Do we count a player from Achaea or Simutronics as worth more than a player from, say, Everquest, because the text players pay more? Do we count the subscribers from Runescape as half or a third of a player because they're paying so little?
--matt
Posted by: Matt Mihaly | Aug 12, 2004 at 12:39
The PC-Baangs mess up the numbers because it means there's no correlation between the number of accounts and number of players, since most of those accounts are duplicates for people who play from several different Baangs. It also seems to be easier to start a new account than to restart characters on old ones, which further munges the data.
--Dave
Posted by: Dave Rickey | Aug 12, 2004 at 12:48
To say there is no correlation is not correct, I think. There is certainly SOME correlation, just like there's SOME correlation between # of accounts on Everquest and # of players. In neither case is it a 1:1 correlation and in neither case do we actually know what the ratio is.
Posted by: Matt Mihaly | Aug 12, 2004 at 12:57
With EQ you can do some surveys asking about the numbers of accounts per player and make a good guess at the actual number of players. With Lineage, most of the players wouldn't know. Also, at least there is a direct and fairly inflexible correlation between the number of EQ accounts and the amount of money SOE makes, where the same is not true with Lineage and NCSoft. And where we can reasonably assume the account to player ratio in EQ is between 1.2 and 2.0 (1.6, +/- 25%), that of Lineage couldn't be estimated closer than 4.0 to 16.0 (10.0, +/- 60%).
At any rate, revenue numbers are probably more significant for business purposes than anything else (I realize that academics would have different priorities, and Lineage's cultural effects in Korea are far more significant than those of EQ in the US/Europe). By that measure, M:TGo is probably bigger than AC, even though it has a much smaller playerbase.
--Dave
Posted by: Dave Rickey | Aug 12, 2004 at 13:56
Well... yeah. :)
Posted by: Jeff Freeman | Aug 12, 2004 at 17:09
The chart is great!
Regarding user numbers,
Investment Analysts tracking NCSoft and other listed MMO developers use either peak concurrent users or average concurrrent users as the measure of user numbers. However, as APRU in China is around $4 per month vs. $15 in th US, they also use total revenues and APRU.
So, based on peak concurrent user, this what China subscription looked like in April (JP Morgan)
Legend of Mir II, 440,000
Westward Journey II, 250,000
MU, 250,000
Legend of Mir III, 200,000
World of Legend (Legend of Mir II var), 240,000
Cross Gate, 180,000
Ragnarok, 150,000
Fantasy Westward Journey, 100,000
Legend of Knights Online, 120,000
Lineage, 30,000
Knight Online, 20,000
There are two versions of Legend of Mir II (#1 & #5) due to breakup of license partners, so if you add it up, the concurrent peak is over 700,000. They had a press release about this milestone a while back before the breakup.
Frank
Posted by: magicback | Aug 12, 2004 at 17:21
Wow. It sounds like Legends of Mir II has easily broken the million mark then in terms of paying players.
Posted by: Matt Mihaly | Aug 12, 2004 at 17:33
Doc, Bruce's site hasn't been updated since March 04... and most of the data (including the FFXI hitting 500,000) is from back in January.
It's a great resource, I wish he would update it!
Posted by: Detritus | Aug 12, 2004 at 20:24
Oh, the above concurrent user chart just give reasons to why Sega is spending lots of money on Shemue. Perhaps that budget includes the cost of their new Shanghai office.
And the news of Runescape reaching 100,000+ subscribes give resons to why more flexible payment models are desired by many players around the world.
A group of friends in China, Korea, or even Spain can go MMOing as easy as going to the movies: they go their local internet cafe, buy some stored value cards, get online together, and log off in a couple of hours.
Frank
Posted by: magicback | Aug 12, 2004 at 20:47
Frank, can you point us to where you got those numbers? Those concurrent user numbers blow everything in US/Europe out of the water.
Posted by: Edward Castronova | Aug 12, 2004 at 22:32
Concurrent users is NOT a good metric because it is hugely affected by play session length. Do the thought epxeriment of a 100 user game where the average play session is 5 minutes versus a 10 user game where the average play session is 24 hours. Most likely the latter game will show a higher concurrency number even though it has 1/10th the user base.
We've seen at SOE that the shorter typical play session lengths of SWG and Planetside make them have lower tie ratios (peak concurrency over subscribers) and lower peak concurrency than EQ, for example. In the case of SWG, we actually get a higher percentage of the userbase playing each day than EQ, but each player pays for less time, so the tie ratio is much lower.
I much prefer daily unique logins as a measure of activity. Revenue is also an interesting metric, if you can get it. :)
Posted by: Raph | Aug 12, 2004 at 22:39
> Here, for example, is Terra Nova's datapage. Not sure why we grew in March and shrunk in April, but we did.
No, we didn't. If you look closer on the chart, it actually shows traffic for ALL of blogs.com. In the table below you can find that terranova.blogs.com is only getting 2% of that traffic. So I don't think there is much correlation between that chart and actual Terranova traffic.
Posted by: Tobold | Aug 13, 2004 at 02:11
A line with aggregated numbers ( Bruce Sterling Woodcock) would be nice, but that is what we can use the excel-sheet for, right?
About the traffic: It is cool for games not public yet, but for the games running at lot more interesting would be the traffic running over the game servers. Anybody who can tell us some IP-Numbers?
Posted by: neven | Aug 13, 2004 at 03:43
The title of the report from JP Morgan is:
China Internet Sector, 24 May 2004 by Dick Wei.
I got the report off the Investext database from my local business library. There is a wealth of investment-perspective figures on the Asian internet sectors.
The headline figures are 250m Chinese middle class population with about 90m internet users growing to about 140m in 2006. About 10m users play online games and this number is expected to double by 20m by 2006. The are mostly on the coastal cities of China. So as they are on the same time zone concurrent user numbers will naturally be higher than the US. Roughly, I would halved the number to get a sort of apples-to-apples comparison with the US market.
I sent some report to Edward, so maybe he'll post some of his insights.
Frank
Posted by: magicback | Aug 13, 2004 at 11:00
BTW,
Shanda Networks, the US-listed Chinese operator of online games with about 7 MMORPGs and 4 casual online games in their operating portfolio had a peak concurrent user of 1,419,000 in 1Q of 2004. That's a lot of player!
Another interesting figure: they have 8,817 servers in 2003 and are forecasted to have 18,445 servers by 2006. Wow!
Source: Shanda, growth priced in, 25 June 2004, Wallace Cheung CFA, DBS Vickers Securities
Posted by: magicback | Aug 13, 2004 at 11:23
they have 8,817 servers in 2003 and are forecasted to have 18,445 servers by 2006. Wow!
Is this growth paralleled in single-player games? What I'm fishing at - is there a different market dynamic in China which favors MMOGs over single-player titles. E.g. software piracy (MMOGs being relatively immune).
Posted by: Nathan Combs | Aug 13, 2004 at 13:00
There isn't much of a single-player game market in China, as I understand it. There's virtually no console market (only two legally approved PS2 titles, is what I heard) and piracy of PC titles is beyond rampant. China has even had cases of pirated MMO servers, much less standalone games.
I'm going to ChinaJoy this year, and hope to learn more. :)
Posted by: Raph | Aug 13, 2004 at 13:29
Nathan wrote: Is this growth paralleled in single-player games? What I'm fishing at - is there a different market dynamic in China which favors MMOGs over single-player titles. E.g. software piracy (MMOGs being relatively immune).
Actually, while muds aren't pirated to the same extent as single player games, piracy of online games is definitely a recognized problem in Asia. http://www.msnbc.msn.com/id/4969821/
--matt
Posted by: Matt Mihaly | Aug 13, 2004 at 13:48
I JUST updated my chart today.
http://pw1.netcom.com/~sirbruce/Subscriptions.html
Bruce
Posted by: SirBruce | Aug 13, 2004 at 23:42
Tobold > No, we didn't. If you look closer on the chart, it actually shows traffic for ALL of blogs.com. In the table below you can find that terranova.blogs.com is only getting 2% of that traffic. So I don't think there is much correlation between that chart and actual Terranova traffic.
That is correct, Alexa is only considering the traffic and rating for second level domains and not third level domains.
That was also the reason why I had to limit the selection to sites using a second level domain. Everquest for example cannot be tracked as it uses a third level domain.
An interesting other measure of the success of a MMOG can be the number of game copies sold. That rate could be extrapolated through the ranking Amazon is doing for every product sold. There would be no absolute values but it could be possible to make an analysis between different MMOGs with several limitations.
Posted by: Luca Girardo | Aug 14, 2004 at 08:39
Raph> In the case of SWG, we actually get a higher percentage of the userbase playing each day than EQ, but each player pays for less time, so the tie ratio is much lower.
I can figure out that Planetside being a FPS game has as consequence shorter gaming sessions but frequenter logins then EQ. But why does SWG has shorter but frequenter gaming sessions then EQ? I fail to see any reason for that. Or does it mean that due to the nature of SWG (game design and “Star Wars” environment), it attracts more casual players ? But then would it not mean also lower login frequency?
Posted by: Luca Girardo | Aug 15, 2004 at 03:28
> why does SWG has shorter but frequenter gaming sessions then EQ?
Well, I suppose because we intentionally tried to design it that way. The much-decreied mission terminals allow people to play relatively short sessions and still advance. The offline crafting and harvesting features allow people to stay offline and still make progress on large orders. I would guess both of those are major factors.
It certainly is an interesting phenomenon, because it's not off by a little--it's off by a LOT. You know the rule of thumb about multiplying peak users by four or five to get subs? It's off by a factor of two or more from that. And yet, each week, a higher percentage of the subscriber base logs in than to any other SOE game--again, by a fair amount, 10% or more to the runner-up.
It may be that we are hitting more casual players, I don't know. This is the sort of data that it is actually somewhat difficult to mine. When we did a poll, the estimated hours played per week came back right in line with other games, so players (at least the polled ones) do not perceive themselves as doing anything differently. Yet the stats show a big difference.
Posted by: Raph | Aug 15, 2004 at 16:22
Yeah, like Raph said, the peak to subs ratio really varies betwen games. I've seen anywhere from 10% to 30%. Although it does seem to stay constant for a given MMOG, I've noted an anectdotal correlation that small MMOGs tend to be closer to 10% and the big MMOGs then to be closer to 30%. Still, if Raph is saying SWG has a higher ratio than EQ... 35%? 40%?... then perhaps the ratio is even more malleable. (According to press releases, EQ's peak to subs ratio is in the 25% - 30% range.)
Bruce
Posted by: SirBruce | Aug 20, 2004 at 05:37