Raph’s Website » Avatar-Based Marketing





By N2H

blog · games · writing · music · art · shop · about Raph · contact

Avatar-Based Marketing

June 9th, 2006 (Visited 6225 times) Tags:

There’s an exhaustively researched article over at the Harvard Business Review on this topic.

Among the tidbits they drop are some Second Life stats:

165,000 total active residents,
of which 65,000 are paying in some form,
and “over 3000″ are making real world money,
averaging $20,000 a year each (skewed by the really high earners, of course).

You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.

32 Responses to “Avatar-Based Marketing”

Jump to reader comments » | Leave a reply »

Trackbacks & Pingbacks

Faith wrote on June 10th, 2006 at 11:37 am:

[...] Comments [...]

Avatar-Based Marketing on Harvard Business Review - MMOz wrote on June 12th, 2006 at 6:59 am:

[...] Avatar-Based Marketing on Harvard Business Review Avatar-Based Marketing on Harvard Business Review Quote: [...]

The Click Heard Round the World wrote on June 12th, 2006 at 8:09 am:

Second Life Statistics Debates

My posting about Second Life reaching the “quarter million residents” figure has generated quite a lot of discussion and debate about the meaning, if any, of this statistic. On the Second Life forums [viewable by registered SL residents only], most

New World Notes: QUARTER MILLION, PLUS CHANGE wrote on June 12th, 2006 at 9:09 am:

[...] Lots of debate on this going on elsewhere too. [...]

Second Thoughts: Harvitar and Groupitar wrote on July 18th, 2006 at 2:32 am:

[...] OK. So he got all the media/blog hits he wanted, SL and RL, and got the ultimate pay-off from Raph Koster, who called his article "exhaustively researched". [...]

Reader Comments

SirBruce said on June 9th, 2006 at 2:10 pm:
Wow! What a timely link… I’ve been talking with Linden for the past couple of weeks trying to get them to provide a number for either “monthly active uniques” or “revenue generating accounts” (subs + people buying linden dollars).
Raph said on June 9th, 2006 at 3:18 pm:

I do wonder how they define active residents (Cory, you reading?). Trailing how many days? I’ve read 60 days elsewhere.

The peak concurrency is around 6000 or 7000 I believe, which matches a tie ratio of 10x for active paying, which is what we are used to seeing in the more casual games. But that would suggest that the other 100,000 aren’t very “real.” Maybe they have an even lower tie ratio.
Prokofy Neva said on June 9th, 2006 at 4:45 pm:
It’s not Harvard Business Review that “researched” this BTW. They just cut and pasted it from the presentation that Philip and Cory Linden themselves did for Google Tech Talks back in March, which is available publicly as a video here http://video.google.com/videoplay?docid=-5182759758975402950

It’s funny how the blogosphere didn’t catch up with this right away, and when they did, they didn’t really subject it to much questioning. If you keep a watch on the “residents online” number even anecdotally, you see that yes, it’s 6000-7000, up from 4000-5000 not so very long ago. How can you tell the real numbers? Perhaps you have a magic formula how to extra the real numbers out of Residents: 245,592 Online Now: 6,188 that I see this second.

I’ll tell you my own magic formula. When Philip Linden replied to my query about what “picks people pick,” i.e. did a data base query on the picks people put on their avatars as their favourite spots (a query I thought would be an important democratic corrective to the gamed and traffic-driven “Popular Places” which beget more “popular places,”) his query came back with like 30,000 people who had bothered to fill out picks. I find many people don’t fill out their alts picks, just their mains. And they don’t bother to fill out picks if they are chance players. So 30,000 pick-pickers is as good a number as any for me.
Prokofy Neva said on June 9th, 2006 at 4:47 pm:
Also this stuff about the “3,000 earners” (elsewhere Philip has said “2,000″) is subject to intense query, I hope. If one person makes $150,000, as we have already discussed to death, and the other 2,999 make only $15,000 a year, well…Not sure how a “living wage” is defined, but I have tenants who easily make $100 USD take-home a day from their content. Of course the average person makes nothing like that, and is happy just to pay tier or pay $40 for monthly entertainment like a cable bill.
Raph said on June 9th, 2006 at 5:06 pm:

The Second Life bit alone isn’t the part that made me say “exhaustively researched” — this is one of the few articles I’ve seen that actually acknowledges the breadth of the space and the diversity of worlds to be found in it. It displays a better knowledge of the space than most of the investors I’ve spoken to.

I agree that the earners data is hard to verify.

The “tie ratio” is the magic formula, but it’s rather inexact.
Over00 said on June 9th, 2006 at 5:47 pm:
Wish I had time for a “second life”. I guess I have too much hobbies for that
Jim said on June 9th, 2006 at 7:40 pm:
Ok… 20k * 3000+ = 60M+… that’s a sizeable economy there.

If we assume Raph’s right, and they only have 60k “real” actives, those actives are spending over $1000 per year (on average) for their Second Life. Yow.

Taking the other end of the range (2k * 20k = 40M, 160k subscribers) sounds more reasonable, you still them dumping $250 or so a year into the game, or about the cost of two subscriptions. Not so unreasonable.
Morgan said on June 9th, 2006 at 10:04 pm:
Raph, I sent you the PDF of this article a month ago. Did you get it? The online reprint doesn’t contain any of the tables or cool illustrations.
Raph said on June 9th, 2006 at 10:14 pm:

Yeah, I got it and read it back then, but then I saw the link to the webified version today and decided to blog it.
SirBruce said on June 10th, 2006 at 2:09 am:
Hey I wouldn’t mind a copy of that PDF too, Morgan, if you still have it to give.

I’m also convinced that a big chunk of SL users aren’t real. According to their economic stats page (you have to have an account to see it), although they have 248,924 users, onle 121,066 have logged in the past 60 days.
helpfulharry said on June 10th, 2006 at 5:56 am:
SirBruce, you should set up a wiki on your website so people can chime in with references to primary sources, for example:

http://planet0.planet-zero.org/~chromal/slpop/

We couuld also point out that the top 2% got developer incentives, and the last time that happened was to 164 avatars in 04-18-2006, which means there were 50*164 actively paying avatars to LL, assuming they were staying true to the ratio on that last payment. Kinda freaky, no?

Well, the problem is, that Anshe is now adding islands faster than LL themselves, and that’s because people are paying Anshe who is paying LL.

However, we also need to look at total L$ of 701M. Given that we assume there are 4500 L$ per user (which is Philip’s magic number), that means we have about 155K active participants in the SL economy.

It’s a complicated picture. To go on actively paying subscribers I think is a huge mistake (SL has more concurrent users than active subscribers, probably)

You also have to think of things like how many servers (probably ~1500) servers that SL has. That’s a *huge* amount of servers, and certainly puts it in the 200K user count MMOs. On content alone, SL is by far and away the largest MMO in terms of 3d content / textures / etc.

And then there is the press. Cover story on businessweek, article in the economist – we’re not talking eve online here.

Anyways, there is a lot of detail that you are missing out on and coule probably get if you threw up a wiki that we could all participate on.
Raph said on June 10th, 2006 at 7:27 am:

Given that we assume there are 4500 L$ per user (which is Philip’s magic number), that means we have about 155K active participants in the SL economy.

Doesn’t it just mean that there’s 155k people with accounts who have $4500 hoarded? It doesn’t imply “active participant in the economy” if they are not transacting. I mean, my accounts have L$ and I have never bought or sold anything.

You also have to think of things like how many servers (probably ~1500) servers that SL has. That’s a *huge* amount of servers, and certainly puts it in the 200K user count MMOs.

No, this figure is meaningless without a strong ratio of avatars to servers. EVE is almost certainly a larger game, with less servers, for example. Having a lot of infrastructure with fewer users doesn’t “put it in the 200K user count” range.

On content alone, SL is by far and away the largest MMO in terms of 3d content / textures / etc.

No question it’s the largest in terms of sheer volume of individual assets, yes. I am pretty sure that no audience would consider it the “best” however, and also pretty sure that even SL residents would discount 90% of the stuff there as crap. I’m definitely a believer in the user content model though.
SirBruce said on June 10th, 2006 at 3:38 pm:
We couuld also point out that the top 2% got developer incentives, and the last time that happened was to 164 avatars in 04-18-2006, which means there were 50*164 actively paying avatars to LL, assuming they were staying true to the ratio on that last payment. Kinda freaky, no?

The top 2% what? Certainly no subscribers, so I don’t know that that means.

It’s a complicated picture. To go on actively paying subscribers I think is a huge mistake (SL has more concurrent users than active subscribers, probably)

I’m sure it does. So does RuneScape. You don’t want me counting all of RuneScape’s free subscribers.

As Raph pointed out in his earlier blog entry, it depends on what you’re trying to track. Since I’m primarily tracking subscription MMOGs, the only way to compare SL to them is on some count of revenue-generating users. I’m happy to track other stats, too, but for the purpose of comparing them on the charts, I don’t think there’s much choice. At some point, maybe all games will provide ACU, PCU, and monthly unique numbers as well, and I can start tracking those.
Wudu said on June 11th, 2006 at 7:24 am:
Does anyone know of any games track player hours? The total of hours played per sub? Woudln’t this be a good measure of a games usage? How bad would AFK players affect this number?
Raph said on June 11th, 2006 at 8:55 am:

I think all the games track it (albeit imperfectly — there can be weird wrinkles in terms of things like whether you count people at character creation or selection, especially if there’s a chat space there) — whether they make it public is another story.
SirBruce said on June 11th, 2006 at 12:00 pm:
It would be nice to have all those numbers. Okay, so many companies are worried about losing a competitive advantage. How about giving researchers access to all of those numbers from 1996 – 2004? That would be a treasure trove of information, for sociologists, for economists, for investors, and for future MMOG developers.
Morgan said on June 11th, 2006 at 2:53 pm:
Bruce, you can get the PDF from Darniaq or Raph. I think I sent the file to Darniaq, but that may have been another article… I’m in the beautifully vast and open desert state of Idaho now, so unless you’re not in a big rush, I can’t send you the PDF until I get back on June 14.
magicback (frank) said on June 11th, 2006 at 7:46 pm:
On SL subscription count:

How many people are renting servers and what is the total revenue?
SirBruce said on June 12th, 2006 at 3:32 am:
I’m not in a rush, but if Raph reads this he can send me the PDF.
Prokofy Neva said on June 12th, 2006 at 5:04 am:
>We couuld also point out that the top 2% got developer incentives, and the last time that happened was to 164 avatars in 04-18-2006, which means there were 50*164 actively paying avatars to LL, assuming they were staying true to the ratio on that last payment. Kinda freaky, no?

This isn’t the top 2 percent of all subscribers but only the top 2 percent of *landowners* who generated traffic on their properties. The overwhelming majority of residents are non-land-owners on basic accounts. This developers’ incentive program was discontinued in April, those are the last figures for the program. It’s hard to know how many total landowners there are, or how many premium accounts there are with a minimum of free 512 m2 tier, because when we were drilling on these figures in the past, we found that a year ago, when the population was 40,000, the Lindens said there were 9,000 premium accounts of which only 6,000 used their 512 tier or tiered up higher for land ownership. People bought those accounts for the weekly stipend of $500 a week which had greater value a year ago.

What is the number today? hard to say, as the incentive for buying a premium account at $9.95/month to get $2000 Lindens is gone, now that $9.95 buys $3000 Lindens, and even looking at discounted accounts for quarterly or yearly (which few get), it’s hard to justify given that increasingly people are renting from islands or mainland dealers rather than buying mainland.

If 2 percent of the top were given the award, that should help generate the totals, but it doesn’t because it seems as if they don’t include the 512s unused, or used only on first land — for some reason LL has been reluctant to release any of these figures (i.e. the true number of premium accounts).

The economic statistics page shows currently 1,012 islands owned (there are some 2500 servers). Each one of them pays $1250 to buy the island, and $195 a month in maintenance fees after that.

Those who buy mainland servers pay anywhere from $1000-$1200-$1500 on the auctions,and $195/month after that.
Allen Sligar said on June 12th, 2006 at 8:50 am:
Wouldnt mind having the PDF either Raph
Raph said on June 12th, 2006 at 8:58 am:

Doesn’t look like I have it anymore, sorry!
Hamlet Au said on June 12th, 2006 at 11:23 am:
Just update my take on this conversation. Bruce, specifically to the figures you’re looking for in your MMO figures, here’s a particularly key point you may want to factor in:

http://nwn.blogs.com/nwn/2006/06/quarter_million.html

[H]ere’s another number to toss into this mix: $5 million. That’s the estimated value of monthly user-to-user transactions last January, as reported by BusinessWeek in May. It’s surely much larger now, but taking $5 million as a base rate, this means 250,000 SL residents are paying at least an average of $20 per month for their in-world activity. This is $20/avg. that residents pay each other, and doesn’t even factor in revenue paid to Linden Lab through land use fees, etc…
SirBruce said on June 12th, 2006 at 4:06 pm:
Thanks for the heads up. Obviously many of those 250,000 are never coming back, so it’s not really a fair comparison. And average revenue per user tells you something completely different. Obviously a game where 10,000 people pay $1,000 each to play is nowhere near as influential to the market as a game where 1,000,000 people pay $10 each to play, even though the dollar amounts are the same.
Elle Pollack said on June 12th, 2006 at 10:37 pm:
A median figure for the amount people earn in SL would probably be more usefull than an average. For comparison, I’m way at the bottom of the statistic pool: I have a basic account and I sell a few avatars in a group store. Between that and the occasional comission work, I make between L$3000 and L$5000 a month. So, pocket change really.

I want to say that Philip has a chart of this; I could swear that his presentation at GDC this year had a graph of this very statistic. But it’s not in the copy of his slides that are on the GDC website.
Prokofy Neva said on June 18th, 2006 at 8:02 am:
his means 250,000 SL residents are paying at least an average of $20 per month for their in-world activity. This is $20/avg. that residents pay each other, and doesn’t even factor in revenue paid to Linden Lab through land use fees, etc…

Hamlet, what we need to pin down about this figure is whether it reflects BOTH payments to “Currency Linden” to post on the LindEX for sale to USD, and cash received in Lindens from purchasing Lindens in USD-denominated amounts from credit cards, or whether it *only* includes inworld transactions. It’s also important to figure out whether they include all those “sources,” like the $500 a week in stipends, and the “sinks,” like the $10 per texture upload, the $100 for group formation etc which is burned and doesn’t stay in the economy. I’ve never gotten a clear answer about that despite asking many times.

To be a fair portrait of what the world’s economy generates, it seems to me it has to leave include LindEx sales, stipend deliveries, and sinks in fees — or else eliminate all of those.
Raph said on June 18th, 2006 at 10:30 am:

I have to point out that if we’re talking about user-to-user transactions, any world will seem to have a huge amount, and most of the big games will show up as much larger than SL. That’s what Ted Castronova used to initially estimate the EQ economy… user-to-user transactions are not qualitatively different just because of the original source of the currency.

Recent Trackbacks

Twitter @raphkoster

Nice 3 mile walk with my daughter, except for the heat of the day. Down the hill, around the park, sushi for lunch, then back up the hill. 17 hrs ago
Rereading a transcript of my late paternal grandfather telling stories: Armistice Day, about Pearl Harbor, about buttermilk in milk bottles. 1 day ago
The world has gone mad. http://bit.ly/vgyE7 1 day ago
More updates...