New Year Survey: How Many Active Second Life Users Will There Be in 2010? How About Maximum Concurrent Users?

Online Surveys & Market Research

Last January, I asked New World Notes readers how many active Second Life users there would be by the end of 2009, and most of those who took the survey guessed incorrectly: over 50% said the numbers would be 700,000 or less. Just 14% who took the survey (and Cory Ondrejka) guessed right, with 750,000. (To complicate matters, the definition of "active user" was tweaked by the Lindens after that prediction, once designated as users in-world monthly for over an hour, whereas now they're defined as unique users who go in-world at least twice a month.) In any case, this still remains the most the important metric for Second Life's future health and prospects, so I'm asking it again this year.

About as important is Second Life's maximum concurrency (the number of simultaneously logged in users), which peaked at 88,000 last March but has been slowly declining since then, to about 75,000. So let's estimate the 2010 status of that too:

Online Surveys & Market Research

Discuss your answers in Comments!

Comments

[ColeMarie Soleil]

wow both of my votes were in the top percentile for both areas XD
not a good thing?

[jjccc Coronet]

this survey is a waste of time it should be a survey that say will Linden Labs fix the Lag yes no my answer is no.
will linden labs fix the constant crashing we all no the answer to that one NO

[Wizard Gynoid]

In my view, this definition of Active Users is not useful. Unique users logging in at least twice a month? That doesn't sound very active to me. There is a core of Hardcore Regular Users. They know who they are. That group needs to be more accurately defined and tracked. I suspect that the number is something like 250,000. And that is a far cry from the "millions" we keep hearing about.

[Ciaran Laval]

Both will rise, I think this will be a good year for Linden Lab, it's 2011 I'm concerned about.

[Tateru Nino]

Insufficient data for anything more sold than coin-flipping.

Are you sure about the change in definition of active users -- because apparently it didn't cause any change in the actual numbers.

[Rusalka Writer]

The problem for me is that there is a large and growing gap between what Second Life looks like and what every other video game in the world looks like. Yes, I know that there are issues with server side/client side systems, but SL hasn't looked any better for a long, long time. I run a couple of businesses in SL and do well, but I don't go there just to spend time or explore any more. Exile or ban most of what people used to do for fun in-world, don't protect creators adequately from abusive exploits, and I just see it shrinking.

[Tymmerie Thorne]

I answered the survey. But I am having an aneurysm looking at it. Hamlet, for the love of all that is good, please put your categories in rank order in surveys/polls/etc.

[Hamlet Au]

!!! Thanks for taking it, Tym. I thought it would make for more thoughtful choices if it was displayed randomly.

[Banana's Hubby]

yes i agree with Wizard. the definition of active users is not useful. i wouldnt call someone that logs in 4 or 5 times a month an active user!
and btw i would love to see a comparison between the increase in registered accounts and the increase in active users. that would explain a lot of things about how users join sl and then leave after a week.

[Valiant Westland]

I have to agree with two concerns already expressed. First, defining "active" as twice a month activity would mean that I am also an "active" car washer and shrimp eater!

LL's definition of "active" points to the disconnect between the recreational and business/professional use of SL. "Active" in the context of "how often do you play __________ online game" might be appropriate. If you consider SL a technology platform, like WebEx or Skype (Collaboration), PowerPoint or BlackBaud (Training/Education) and PhotoShop or Autocad (Graphics/Design, twice a week use is far from active!

Secondly, polling isn't about style or fun or randomness. The goal of any poll should be clarity, brevity and above all, accuracy.

I would like to see someone in the academic community, perhaps a statistically minded person like Beyers Sellers, do a more scientific analysis of the use of SL, compared with both other MMOG environments as well as professional technology tools and present these results on Metanomics (Dusan are you there?...)

[LittleLostLinden]

Define concurrent users?

Is that concurrent users meaning real live people, or concurrent users meaning real live people including bots, campers, and model avatars?

http://pixelscoop.net/tag/second-life-traffic/

[Hamlet Au]

"In my view, this definition of Active Users is not useful. Unique users logging in at least twice a month? That doesn't sound very active to me."

Wiz, monthly uniques is the industry standard, so it's a better metric for comparing SL activity to other virtual worlds/MMOs. You're right that we can look at user hours, but I don't think it should be one or the other. Facebook reports YoVille has 20M active users, but that includes folks who log in for 15 minutes once a month; same with Habbo. Why should SL be held to a different standard?

[Wizard Gynoid]

My good friend Jopsy Pendragon has a formula for Hardcore Users. "I really expect hardcore users to be approximately 3x-4x max daily concurrency."

[AnnOtooleInSL]

This survey assumes no major leadership changes will be announced for Linden Lab in 2010 right?

I suspect some changes will be made and things will look better mid year. About the time the Soviet KGB expert's predictions the USA will disintegrate and Linden Lab becomes owned by China (at which time SL will be closed and Hipihi takes over right? lol) . http://online.wsj.com/article/SB123051100709638419.html

And people think my predictions are whacked lmao.

Something has to change big time or all of the numbers are way too high and median concurrency will probably fall below 35,000 if not outright approach zero as the domino effect kicks in. (Current median average is around 50k to 70k so maybe people need to be looking at the realistic averages instead of peaks to get a clue as to how things are going.

I.e.; the trend is on a solid down trajectory. A trend that commenced around March/April 2009. Someone said the trend commenced the day the anti adult content position was announced by LL.

This trend, if it stays on track, and linear, would arrive at zero in only 3 years. But once it passes a certain point the drop would increase dramatically as the population vacates to staunch losses flow. So on the current trajectory under Kingdon's lead we have maybe a year and a half left. So enjoy it while you can. It is pretty obvious LL isn't listening to customers.

I would love to be proven wrong.

[soror nishi]

Yep, have to agree with Wizzy on this. It doesn't matter how StupidVille players are recorded, twice a month is not an active user. That's pretty obvious to me. So...we are talking about a stupid statistic, and... honestly, why bother.

Daily concurrent usage I can monitor myself without LL tampering or definition change. The fact that we are losing a good deal of the 15000 or so bots will alter those figures. I am personally not that interested in 2 million users. I see 40 to 70k online. That's a figure that would be far better to use as a discussion point, in my opinion.

[Hamlet Au]

Soror, by that logic, if a musician logs into Second Life twice a month to perform for an hour, she doesn't count as an active user. Sorry, they seem active to me. For that matter, on occasional months when I'm really busy with other projects, I log into SL about that much. In any case, that definition of "active user" is what folks in Silicon Valley are looking at. When and if that number starts going up in a big way, you'll start seeing a serious return to SL.

[LittleLostLinden]

I don't think your Peak Concurrent User data of 75,000 users is valid because it does not take into account bots\campers\model avatars:

http://spreadsheets.google.com/pub?key=t_J-MGGwJhhLL8mNUo_9vyQ&output=html

Perhaps if you reword it somehow so it states that LL's numbers suggest 75,000 for 2009.

Everyone knows the true number is much less than 75,000 users.

A user to me is one person, not one user account.

One person can easily have 40 users accounts.

Perhaps if you reword it to say 75,000 user accounts as to make the ditinction between users and user accounts.

[Hamlet Au]

LLL, that link (even if accurate) only counts a few hundred bots. How many bots are you claiming there still are in SL, despite the ban, and what's your source?

[Hamlet Au]

LLL, that link (even if accurate) only counts a few hundred bots. How many bots are you claiming there still are in SL, despite the ban, and what's your source?

[Fogwoman Gray]

Considering the less than auspicious start to the year inworld, I am afraid that I am predicting a decrease. The infrastructure obviously cannot handle even the current numbers, much less any increase. Asset server issues are increasing in frequency, crashes, downtime. Basic functionality is being lost. They are frantically trying to mend and patch just to keep the world online, inventory losses are increasing outside of the steps LL is actively taking to remove some content. My suspicion is that the lab is just trying to keep the current platform limping along until 2.0 is ready. The question is, will it come before everyone has given up in disgust.

[LittleLostLinden]

The source is myself. I was frustrated one day while constantly seeing more and more bots\campers\AFK model avatars inworld so I started compiling the list. I plan to add photos soon.

I'm not certain how many I would claim as one of the above, because as you mentioned, there is only about 400 on the list, but this is a list that was drawn up from random locations in less than 2 hours.

So, if I can easily randomly find 400 in that short of time, one could make a good guess that there are several thousand still inworld, since the ban was invoked.

And the ban, really isn't even a ban at all if you think about. There is no way to report bots\campers yet.

If you attempt to report them inworld for instance, by using the Report Abuse feature, you will be told that LL is taking care of the problem by monitoring search.

My question is, if that is true, than how is it that so many are falling through the cracks of the Linden Labs bot\camper ban?

I would say there are at least 10,000 to 15,000 bots\campers inworld, probably more. I'm just starting this research but one thing is clear.

Since the supposed ban of bots and campers, the numbers of them are still very very high indeed, and it is still a major problem.

So, when you quote that 75,000 users number, well, it is far from accurate because you are including bots\campers\AFK model avatars in your number.

[Hamlet Au]

I'm very skeptical there's still anything like 10-15K bots extant, LLL, but I'm willing to look into it. What's your SL name, for attribution and getting in touch with you in-world? Feel free to contact me in SL or by email.

[LittleLostLinden]

I have it posted on the site.

[email protected]

http://spreadsheets.google.com/pub?key=t_J-MGGwJhhLL8mNUo_9vyQ&output=html

[Hamlet Au]

SL name, please?

[AnnOtooleInSL]

"» 11594 bots registered, 1257 bots logged in during the last 24 hours, 2517 logged in during the last 7 days «
» 1799 bots currently online with a bot-version greater than 3.0.9.5 « "
-- http://www.pikkubot.de/dokuwiki/doku.php

And that is just one bot service provider. There are a lot of bots used one to a parcel for various activities that only a special client can do since scripting won't do it all.

There are still plenty of bots. So instead of peak concurrency start looking at median concurrency. That will remove most of the bot factor in a statistically incorrect manner. But will give a better sense of actual usage.

[Hamlet Au]

Not sure how that follows, Ann.

"if I can easily randomly find 400 in that short of time, one could make a good guess that there are several thousand still inworld, since the ban was invoked"

Even assuming that's the case, several thousand wouldn't be a major factor for the purpose of this prediction. Especially if concurrency sees significant growth in the several tens of thousands actual users (along with vagabond bots). And if there's not significant growth, the bots won't matter.

[LittleLostLinden]

Thanks Ann.

Can you tell me if a similar tool exists for detecting campers and AFK model avatars?

From what I have noticed, it seems as though many former bots have made a shift to become AFK model avatars.

[LittleLostLinden]

Sorry for the double post, I forgot I had a few more questions.

"And that is just one bot service provider."

In your estimate, how many bot service providers are there, and can the bot service providers be tracked easily, or are they constantly changing?

I would like to see some more of those realtime statistics for other bot providers if you have them.

Is there some way to monitor camping software in a similar manner? I know in some sims there is a campmaster or similar product but have no idea what tool could be used to gather the same type of info that your bot tool does.

[Arcadia Codesmith]

Without knowing which bots are registered and which aren't, you can't really make any assumptions about their impact on the stats. You can assume that most of them are unregistered (illegal) dodges to drive traffic, but you can't prove it.

I predict modest to moderate growth this year, despite Linden Lab's continuing efforts to alienate long-term residents. Attitudes amongst current users range from pessimism to rage, but the upswing in the real world economy will be echoed in the virtual world. This will be taken as vindication by Linden Labs for their recent spate of boneheaded decisions, and they will continue in the same vein, because the graphs will look good despite the vibe at street level getting increasingly sour.

In short, I predict we'll stick with this insanely unpredictable platform because none of the alternatives are really viable. Yet. This situation will not last.

[Lalo Telling]

Checking back three days later, to discover more than half of the respondents believe the *best* that will be done is a plateau, if not a decline:

-- 57.5% in the case of "active" users;

-- 56.5% in the case of peak concurrency.

Why am I not surprised?

(It'd be nice to know total number of respondents, too. Will you close the poll and post final results -- and if so, when?)

[Little Lost Linden]

Was there a concurrency report for Q2 2010?