Got numbers?

I read what you wrote before the edit but you failed to convince me :)

The edit was more explanatory. But, excuse me, what is the "conversion rate"? I know the churn rate and the fact that a part of the boxes doesn't even get used, but you don't explain what that third "loss" is.

At any rate I went near with my prevision ;)

I'll also take the occasion to save a message Walt posted long ago:

And, with that said - most of the information listed below is again - only my understanding, and should be taken with a grain of salt. Or a pinch. Or a larger quantity, if you so desire.

Types of MMO users that could be counted - Highest to lowest:

# of Box Sales (+downloads) - This is the highest number for a press release. Fairly easy to verify box sales numbers through the established retail chain.

# of Registered users - This is the 2nd highest number. This is everyone who has ever created an account - through any means. For some reason, a small percentage of users who buy a box never actually install the game and create an account - be it technical reasons, unopened gifts, etc. It is my understanding that the SWG numbers on Bruce's chart use this number.

# of Active accounts - This is the 3rd highest number. Everyone who can currently log into the game and play. Note that this includes players who are still on their free month. This includes players on free trials. This includes players who have pre-paid for time and subsequently cancelled. Adding all of these numbers up and adding in attrition from launch date means that active accounts can be anywhere from 90% of registered users down to 20% (for much older games or games with poor retention). It is my understanding that most of the Everquest numbers on Bruce's chart use this number.

# of Active, paid accounts - This is the 4th highest number. Everyone who is a recurring subscriber and not on a free trial period. This number is roughly 10-15% less than the total active accounts - depending upon box sales and free trials. This is the number that the accountants are most interested in, because this number dictates cash flow, short term and long term. It is my understanding that the WW2OL, DAoC, FFXI, and UO (and maybe some of the EQ) numbers on Bruce's chart use this figure.

# of Active, paid, uncancelled accounts - This is the 5th highest number. For some reason, a healthy (5-10%) percentage of MMO subscribers cancel every month, probably to guarantee that they don't rack up credit card bills for an unplayed game. While no one reports this number, it is worth knowing about.

# of Unique users per day - This is the 6th highest number and a good comparison to the # of active accounts. Roughly 40-50% of the active users will log in in a 24 hour period - so, if you have unique users, you can multiply it by 2 to 2.5 to estimate accounts.

# of Peak users per day - This is the 7th highest number and a good comparison to the # of active accounts. Roughly 20-25% of the active users will be logged in at once, barring some in-game mechanism that encourages you to remain logged in. So, active accounts can be extracted by multiplying peak users by 4 or 5. Currently, only DAoC displays this number publically, to my knowledge, but other games such as EQ (and legend of Mir II, apparently) release this number regularly.

It would help a lot to know the average peak users per day right now...

"Conversion Rate" is the rate at which accounts created become paying customers after their free time runs out. It's basically a judgement on the newbie experience.

Active Subscribers X (1 / Percentage Churn) x (1 / Conversion Rate) = The number of boxes you need to sell each month to maintain a steady population.

If the first set of numbers applies, then Blizzard needs to sell about 63K boxes a month to maintain a 440K population. If the second, then they have to sell about 42K boxes a month to maintan 550K. Basicly, conversion rate, churn, and your sustained sales determine your stable population. Drops in any of them affect the size of your subscriber base. There's a "Matthew Effect" here, on one side of the curve your game grows, on the other it shrinks.

The general effect of new game releases is a one-time increase in your churn rate for the period, and usually a reduction in your sustained monthly sales. And conversion rate tends to drop over time, probably because as your playerbase matures the newbie experience degrades (nobody to group with, less incentive for current players to make friends with newbies).

--Dave

My current estimate for WoW is 1.25 million. We know there's at least 750K from their statement, and given the number of concurrent logins in Europe and South Korea, 1 million is pretty soild. Of course, some of these people are on their free month.

Dave, I disagree with your analysis, because you're basically saying Blizzard is misleading about the 750K figure and is reporting all accounts created (Walt's #2 above) instead of current subscribers. I just don't believe that to be true because that's not what they said. Having said that, that would be claiming a 93.75% total conversion rate, without factoring in any quitters, and I do find that hard to believe. Still, even if your assumptions were true, you still can't ignore the Europe and Korea figures. Can we agree that it's somewhere in the range from 750K - 1.25M worldwide?

Abalieno, as Dave said, you were wrong, although I don't know why my explanation wasn't clear. My estimate was basically the same as yours; we differed by only a little. I was being conservative, of course. I agreed it was at least 500K in North America. Turns out it was 750K, which if true means we were both too low in our estimates. :)

Now, as for that "Walt" post above, I looks vaguely familiar, and I think I posted a response to it. As for "# of Active accounts" vs. "# of Active, paid accounts" I have no real opinion. Some MMOGs report one number and others report the other, and often I don't know who is reporting which. I see nothing wrong with people reporting accounts currently on their free month as a "subscriber" for that month if they choose to do so; it's only really a major factor when a MMOG initially launches. As for the "# of registered users" metric, I do *not* use that, for SWG or any other game. There was an early press release regarding SWG where they used a registered user's number that was the same as a subscriber number data point I used, but I later verified that was roughly accurate for subscribers as well, not just registered users. Again, when a MMOG is first launching, this probably included a lot of folks on their free month.

Bruce

Walt post was on Q23 but he used to paste it everywhere and so did I, so I don't know where you saw it.

By the way, both Blizzard numbers and Dave numbers don't seem convincing, for the reasons you explained. At this point is hard to know who's right. Maybe you can set what Dave said as a floor and what Blizzard said as a roof but the doubts won't vanish.

From the start, more than one year ago, my prevision on WoW was that it would reach at least 400k in six months. That's what I repeated everywhere till the release. Again I was basing my estimation on the numbers of the current market, knowing the quality and type of the game. WoW strongly cannibalized the current scenario, they also introduced many more players but if the total goes above 500k I start to feel doubtful. Still, I've never seen a mmorpg remain at the first places of the selling charts for months. The last week WoW was actually FIRST.

There are all these elements to consider. The concurrent log ins don't matter because they reached that peak during the Christmas and right now they are nowhere that cap (the queues are way reduced with no changes in the infrastructure).

The point is that the difference (about NA) between 450k and 750k isn't that small and both are somewhat justified. Less justified, instead, is a blend between those two numbers, because it would mean that both Dave and Blizzard are wrong. And it isn't believable again.

The numbers I offerred are only a "best guess". But the 750K number *isn't* believable, it assumes not only a 93.75% conversion rate but a 0% churn, sorry, not going to happen. You can plug different numbers into my formula and get different results, arguing that the Blizzard polish is giving a higher conversion rate, and that the Rest system and a flatter levelling curve is giving a better churn, but it's pretty much the formula that has to apply, Blizzard hasn't done anything so revolutionary as to make the model inapplicable. Even with improbably high rates (90% conversion and 4% churn), you get 650K as the upper bound, so I'll stand by my numbers until Blizzard releases something firmer.

I didn't try to factor for Europe because as you said, they're still on their launch bubble there. And they appear to be using the Baang-centric business model in Korea, so there's no box sales to plug into the formula there. As an over-all figure, your's seems close enough for the limited information we have (and in fact, I said that if they don't bust the 1M threshold, they won't miss it by much). I'd tighten your error bars a little bit, putting them somewhere between 800K and 1.1M worldwide after it all shakes out. A lot of that is going to depend on their sustained sales, whether or not a "Waiting for DDO/MEO" effect sets in. If either or both is further delayed, or the beta reports are highly negative, then I'd revise the estimate upwards.

--Dave