May 2007 Key Metrics Published

Hello Everyone!

I have published our Key Metrics – our 6th Monthly edition, now through May 2007. They are available as xls, ods, and Google Doc formats.

There are several changes and expansions we are making to these metrics, to make them more thorough and relevant for you. I hope you find these useful, please continue to provide constructive feedback in the comments and we will keep improving. I’d like to thank you all for the comments and trackbacks from so many of you interested in these figures – obviously this is an important topic for many of you. 🙂

Explanations in changes of the data and layout are provided after the fold below, but here are some quick highlights:

1. Including the Total Counts of Active Users now, not just the percentages. This was highly requested by many of you, both in the blog comments and at my office hours and I agree, it makes these numbers more relevant and credible.

2. Expanding our metrics from the count of active users over to a hopefully more interesting metric – Number of Hours spent inworld. I think this important because of many of the points raised in this article on “The Attention Economy” which I believe applies to our world. You may note variances in the percentages of the Counts of users versus Total Hours spent inworld for Countries, Age Buckets, and reported Gender. I find these variances to be fascinating, and would love to hear how these differences strike you. This topic in particular I would like to focus on for this week’s Inworld Office Hours this Friday, 10 AM at my office in Beaumont.

Thanks for the ongoing conversations, and I hope to see many of you inworld soon!

<more info on changes behind the fold>

Changes in the numbers:

A. Inclusion of <unknown> category for Country and for Age, for Active Users statistics. These are due to a couple of factors – the primary being neither our own registration pages nor the current version of our Registration API enforce valid values to be selected, so we have a very small percentage active residents registering with birthdates thousands of years in the future (welcome back!) and/or blank country information (we welcome our users from “Planet Pluto”!). It is interesting to note that while there may be a significant count of users with invalid information, when looked at in terms of Active Hours spent inworld it is significantly less impactful. I am improving Data Quality ongoing, by associating logged-in IP with known country, and upcoming age verification, and cleanup of obviously invalid registration information. By including these current metrics, you will see in the future as the percentages of these Unknown categories will shrink, and we will aim to eliminate them altogether where possible.

B. Including an <Other> bracket for the Top 100 Countries report – so that we can verify that the Total Count and Total Hours figures for the Active Users by Country and Active Users by Age Bracket reports balance, and so the size of the Long Tail of our users from countries with fewer users can be seen – fyi, we have active users in 336 countries in May 2007. Please note: The User Breakdown by Gender Report is not based on Active Users but is based on all logged-in users and will match Total Hours inworld from the “Total Hours” tab.

C. Explanations of the metrics definitions- I have begun a wiki page at Metrics Glossary to begin defining these metrics and keeping answers to common metrics definition questions at your fingertips, such as this question that I’ve seen and answered several times in the comments- the more places we have these answers centralized, the easier they will be for everyone to find and use. Once this page has been used for a couple of weeks and the definitions expanded apon, I will begin linking directly to these definition from the headers of the reports. Your commentary is very welcome at the Metrics Glossary commentary page.

About Meta Linden

Data Warehousing and Metrics Analysis for the world of Second Life. A Linden Lab employee in California.
This entry was posted in Economy. Bookmark the permalink.

74 Responses to May 2007 Key Metrics Published

  1. Emondrell Raymaker says:

    Excellent, I enjoy seeing SL’s numbers. Good job!

  2. Active by country is fascinating, indeed! Now that pesky premium by country…

  3. Ceeq Laborde says:

    Excellent, thank you very much for listening, great work.

  4. Tony says:

    I’m wondering what this big unknown country is that provides so many avatars!

  5. Ceeq Laborde says:

    @4 :Please refer to the explanation in the post above

  6. Pingback: reBang weblog

  7. Aeper Jie says:

    Wow this proves something.
    13-17 (Teen Grid) 4,112 0.81 % 108,940.08 0.60 %

    We dont mean much to LL. So why cant Teen Grid get a Discount on sims and Prem. We dont own credit cards. so yea. but scarry thought how small we are to you guys.

  8. Buckaroo Mu says:

    Heh – Am I the only one who noticed this, under Age Bands??

    Age, avatar count, total hours
    18-24,138,859 (27.34%), 3,119,442.98 (17.14%)

    Just goes to show – it’s those damned kids and their alts.

  9. Freya Mokusei says:

    @7: 18-24yr olds are allowed on SL.

  10. Aeper Jie says:

    Buckaroo you MG foke own Second life. so stop conplaining and go back to your Cybering. ok? i noticed the Age Bands. thats where i got my info. look at post 6

  11. Revolution Perenti says:

    Great work Meta as always.
    love the work your putting into SL, and listing to us.
    see you soon in sl or at your Office hours

    kisses to my favourite spider linden


  12. Meta,

    I think it might be worthwhile adding the national language next to each country. This would provide an indicator of native languages which could be a very useful metric.

  13. AJ Brooks says:

    I agree with Nobody (jeez, it seem wierd to bne writing that). Language would an excellent metric, especially for those in higher education and for learning purposes.

  14. Buckaroo Mu says:

    Aeper, I was not commenting about the Teen Grid – I was commenting about the vast difference between the number of 18-24 year old Accounts on the Main Grid and the amount of time they spend there, compared to the other age brackets. No offense was meant toward the TG. And I wasn’t complaining, either – just making a note of the striking difference.

  15. Pingback: SecondLife - How To Make Money In Second Life » May metrics - the numerical state of Second Life

  16. Victor Gil says:

    What does “active users” mean?? Thanks

  17. As someone who studies the growth of SL in Brazil, is it possible to get a copy of the Brazilian subset of the numbers?

  18. Pingback: Aussie population update - stable at 10K

  19. Daedalus Young says:

    So, the people with clearly inaccurate birthdates and countries, I believe they are in violation with the SL TOS, as can be read also here:
    And so they are placed on temporary hold, right?

  20. Erasmus Hartunian says:

    Matching active users counts against population by country shows fascinating results. Netherlands tops the list, the only country with more than 1 active user per thousand population (barely: 1.04). The US is not even in the top 10 with only 0.44 per thousand. The top 10 are all European countries.

  21. Katrina Bekkers says:

    The last month was just an alarm bell going off:

    This time, it’s markedly worse on every metric:

    The numbers are dwindling down. Sharply. As in “the worst month for sales in the past semester”. And May is even a 31 days month!

    Well, the writing is on the wall. Probably it has been for quite a while. Who knows if LL has the means (or willpower, depending who you ask) to turn the tide?

    “Wait and see” by customers is another way to call “upcoming death” for a service company like LR. I’m worried.

  22. Rex Cronon says:

    I would like to see an new table added with the following columns:
    -either number of images uploaded or amount paid to upload them
    -nr. people that paid to upload images
    -either number of sound clips uploaded or amount paid to upload them
    -nr. people that paid to upload sounds
    -either number of avatar animations uploaded or amount paid to upload them
    -nr. people that paid to upload avatar animations

  23. Pingback: Worlds-Unlimited » Blog Archive » Second Life: weniger Nutzer - aber intensivere Nutzung

  24. Hello,

    I would like to thank you for listening to our suggestion from the previous month to add “active users” to the statistics.

    We very much welcome this new metric because it provides all of us with more transparency about the REAL users.

    To our surprise the number of active users is only halve the number of residents logged-in during last 30 days. This means that only halve the user logging in do stay for at least one hour continuously.
    (Why do we loose the others? We heard of many people experiencing stability issues and therefore the numbers are not really surprising).

    On the other side we were positively surprised by the number of hours active users use second life. In germany the average usage time ist 36,7 hours per month per active users. Thats less than in the US but a good average of western countries.

    Other german figures are as expected – Germany is still number 2 based on absolut user figures and user time.

    Lets continue our good work on serious applications of second life – because it is a great plattform conbining 3D-displaying and efficient communication capabilities.

    To Linden Labs – please work hard on the voice system – more and more business cases we have developed with our clients require true voice communication. The Beta – Version we tested was really great – hope you will include that on the main grid (very) soon.


  25. Felix Duesenburg says:

    A little twist to be kept in mind when reading the by-country stats: People who don’t submit their true RL data tend to pick a country from the top of the drop-down list, i.e. those beginning with ‘A’ since they’re sorted alphabetically. This explains why e.g. Andorra and Afghanistan appear overrepresented.


  26. Kid Kuhn says:

    What does the “count” on “Gender – With Hours” stand for?

  27. Farallon Greyskin says:

    The thing I found interesting was that under the age bands, the older residents (RL age) spent the most time on line on average per person. 18-24 spent the least time per person and over 45 the most.

    Sort of inverted to what I would have guessed… But maybe younger people are more into “games” and older more into the pure socializing or the running a buisness aspect of SL, or maybe younger people have more numerous distractions on and off line and older are more concentrated of their time on fewer thnigs?

  28. Pingback: GigaOM GigaOM’s Top Ten Most Popular MMOs «

  29. Meta Linden says:

    @ Tony & Victor Gil & Kid Kuhn : all of your questions are answered on the Metrics Glossary that I referenced in the post.

    @ Nobody & Erasmus : Linking this data with Language information, country demographics, broadband penetration rates, etc is *exactly* what I’m hoping the greater community will do, using external data mashups, maybe some Google Maps Mashups, hopefully linking back to and attributing the source. That’s another great topic to discuss on Fridays – how can we extend these?

    @ Felix – good point about people choosing Default values during registration. That could also be a factor of why our *count* of self-registered Males is so high.

    @ Katrina – this is a great question to bring to Zee Linden, who often comes to my inworld office hours, he loves to talk about the overall health of the economy and how the Lindex levels are a measure of it. I’m not enough of an economist to comment on that properly, personally.

    @Daedalus – we *could* place the accounts on hold, per TOS, and handle the flood of support calls again, but those are a Lot of active residents. We’re currently looking into better automated mechanisms to update this info (country especially) without being disruptive to customers or to our support team.

    @Jeremiah – we get a lot of one-off requests for specific subsets of the data – either geographic, or line of business, or some other demographic subset, and at this time, we simply don’t have the resources to supply these right now. Please come join my inworld office hours and let’s try to find some way to expand the overall metrics so you can get the answers you’re all looking for.

    Thanks all for the encouragement and support!

  30. Pingback: Multimediablog » Second Life Userzahlen veröffentlicht

  31. UsagiMusashi says:

    You mean your finally omitting some of the stats are incorrect?

  32. Pingback: Remco + Media » Blog Archive » Gebruik Second Life valt tegen

  33. Pingback: Nieuwe Media

  34. Pingback: » Blog Archive » Eto Nina

  35. Seraph Nephilim says:

    @Meta – regarding the default values — maybe suggest to the web team that where possible do not have a default set — require an explicit choice. And on very long lists, put the most common ones up top so they’re more likely to get picked, rather than someone failing to scroll through & just picking the first value they find. This might help get better data — but I’m sure a User Interface expert might have some better suggestions.

  36. Dirk Felix says:

    Could you also add columns that displays complaints and also issues/failures per hour. This would give us all a better picture of SL. Thanks!!

  37. Celebrity Trollop says:


    Would it be possible to break down by gender and country who is buying things in world and how much they buy?



  38. Pingback: Millions of Us » Blog Archive » How much time do you spend in Second Life?

  39. Pingback: Nexeus Fatale » Blog Archive » Understanding the Innovation Potential of Second Life

  40. Pingback: JN’s weblog » Blog Archive » Ontnuchterende cijfers over Second Life?

  41. Pingback: » Second Life gebruik valt tegen

  42. Forseti Svarog says:

    Meta, thanks for adding attention information.

    Question: Why does active users show up at 507K, but F count and M count show up on “Gender – with Hours” as 1,118K ?

    How are F count and M count being derived? I didn’t see it on the glossary wiki.

  43. Pingback: Barry Woudenberg » Blog archief » # gebruikers Second Life overdreven

  44. Pingback: Tenth Life : SL statistics as percentage of country population

  45. Pingback: Pedro Meya Marty - Second Life Agency : Second Life Agentur » Blog Archive » Die Mai Lindenstatistik

  46. Loydin Tripp says:

    Meta I appreciate any and all numbers but changing the rules of each month makes it almost impossible to track trends accurately. If you decide to change what you track or what method you use can you please tell how you are doing it one month and contrast that method to the months past?

    Also, when you list country percentages how you derive what country the usage is from? Is this by the URL they are coming out of or is it by the account information or some other method?

  47. Pingback: Out to Pasture » Blog Archive » Second Life Stats, May 2007: Older = More

  48. Meta Linden says:

    @Celebrity – that is the sort of thing I’m trying to construct with a new “Logical Inworld Business Model” that I’ve been discussing at my inworld office hrs on Fridays. It would be great to have your input on how to make this work properly.

    @Forseti – as mentioned in the post, “Please note: The User Breakdown by Gender Report is not based on Active Users but is based on all logged-in users and will match Total Hours inworld from the “Total Hours” tab.” I am looking to add a breakdown of Gender for “Active Users” and not just “Logged in Users” in a future release.

    @Loydin – I’ll continue to explain the changes and expansions to these reports as we make them, I believe this month’s numbers can continue to be used to trend against previous months (pct of counts) and ongoing these can be trended forward. The intent is not to remove previous stats that are being trended (counts of active), but to expand them to include alternate views (such as total hrs). As I noted in the “A” paragraph under the changes description, Gender, Age, and Country demographics is currently derived from the information provided by the customer apon registration, though we are currently cleaning up Unknown Country via the IP address used to log into the system.

  49. Pingback: » TwoMenDesign » Second Hype

  50. Pingback: El Estado de Latinoamérica en SL, Mayo 2007 « Aenea’s Second Life

  51. Pingback: Get-Klaus » Second Life

  52. Thank you for these statistics. I quoted the one on users’ countries of origin to advocate translations of SL’s Terms Of Service, Privacy declaration and abova all Community Standards, in a post about the research done by Nottingham University on antisocial behavior in Second Life (see ) . According to this statistics, less than 37% of SL users come from English-speaking countries, and while more than that know English as “lingua franca”, this doesn’t mean that they are able to understand these “legal documents”. And even though they must declare that they have read them and shall abide by them when they create an account, bad habits encouraged by skin-wrapped licenses mean that users often tick the box without actually reading the legal stuff. Let alone if the legal stuff is in an unfamiliar/unknown language.

  53. Loydin Tripp says:

    Meta, thank you for your response. I had hoped to get the following comments and questions in here, in time, to capture your further answers. I hope to come in world during your office hours to get more clarification or conversation but it would probably benefit more Second Life residents to see the answers here for future reference.

    My first comment: I do not believe that a Blog is an appropriate place to release and store this vital information. Blog interface and format lends itself to a diary format not to the dissemination of vital business information. Locating each month is a pain and related information is often missed or miss placed. This metrical data deserves a place on the Second Life web site where, if designed well, it can easily be found and bookmarked. A one stop shop for metrics.

    Not only does it make it easier to locate the information and organize it but it elevates its status and credibility. I assume you do not mean to give the impression that this effort and its outcomes are not important. But delivering metrics links and files in this fashion appears haphazard and attended to in an after thought like fashion. I know you put a lot of work into this, it deserves better and so the do the constituents that consume these figures.

    On that page should be the links to the Blog about that month’s releases, your comments and the comments of others, thats what blogs are great for…

    Also, the Glossary needs to be completed and updated and a link from that possible web page. This month’s statistics points out the varying ways that Linden Labs defines ‘Residents’. “unique”, “active”, “excluding alts” and “registered” are hard to keep track of in the figures you are referring to. And once defined you need to use these terms consistently.

    In all further comments I am referring to the XLS file downloads.

    • Confusion in File and Cell Naming Practice

    When residents download the XLS file option we get them on our systems in this naming convention: stats_200706.xls. This is confusing because you are naming the file by when it is created, not what it is about, which in this case is the metrics for the month May. May is the is the 5th month so the file should be named stats_200705.xls. This could be leading to another confusion… you state at the top of this blog, “…our 6th Monthly edition, now through May 2007.” Unless I am confused, I believe you started the metrics in January, delivered on February 2nd. Is that not 5 editions? Is there one I missed?

    It is common practice to always use the same number of cell columns and to always name the first row with the field names and always use the same names.

    • Confusion in Country Naming and Counting Practices

    Besides the major revelation in this month’s stats that their are 507,844 unique, active residents you have a large “unknown” and “others” categories. These categories did not exist in previous months. “unknown” is 5.19% of the total residents, beating all countries but the top 4. How are we to create trends across months? I think this needs to be fixed and released to be of any use. It is standard practice to create a list of records that is always the same length and then if one of the record’s sample month’s stats are insignificant or below measure then it is simply empty or better yet, set to zero. This way the list is consistently the same length and can be processed in a database more consistently. The cell columns should be the same number and in the same position from month to month, as well as, the aforementioned cell naming practice. In your monthly samples, countries pop in and pop out, vanish and reappear, it cannot be determined if they indeed have been pulled for some political reason of censorship or if they are just fluctuating in interest and use.

    Actually, what would be more helpful, would be, to not assume that everyone owns or wishes to use Microsoft Excel, but to instead deliver a more flexible set of ‘tab delimitated’ text files. Any spreadsheet or database application can open these files without having to convert them. But if the amendments and cleanups were done, I could live with XLS.

    I was happy to see that you returned to the longer and more accurate practice of full decimals places for the country’s share numbers this month as you had in all months except for the April metrics released in May. In that sample the percentage shares are so rounded down to be almost useless for accurate comparison across months. Also, I do aggregate computations of Global Region and Language share, many of the smaller percentage share countries ‘flatten’ to the same numbers, sometimes called a ‘floor’, this makes these aggregates less accurate. This and the practice of placing countries in the category of ‘others’ tends to say to them that they are unimportant and do merit counting, even though small. I am sure this was not your intention. I hope that you will release this version of the metrics with the longer decimals.

    The source that I would recommend to use to determine what’s a country and what is not a country, as well as, the correct name and spelling of the country is the “CIA World FactBook” Home page here: and the very useful and ‘ready-to-go’ tab delimitated text file for download:

    I do not recommend this source because I like ‘spooks’ or because I agree with their somewhat politically slanted views of the origins of countries and their governing style but because in the facts department there is no better, more accurate or up to date source of country information. The World FactBook lists 233 countries as of June 2007. It is relied upon by the US State Department, other federal agencies and many foreign governments.

    They did not include in their list, Guadeloupe or Reunion, I just added them. But if you were to use this list and always stick to it then it would be easier to make better determinations over time and by other details, avoiding ’round off’ errors.

    • Some detailed errata:
    • “Azerbaidjan” is spelled “Azerbaijan”
    • “Bahamas” is “The Bahamas” or in alphabetic list format “Bahamas The”
    • “Bosnia-Herzegovina” is “Bosnia and Herzegovina”
    • I have never heard of the country “Satellite Provider” listed for January, should we ignore?
    • This month you blended the numbers for “South Korea” into one category “Korea South” but in the past you have listed it as “Korea South” and “Korea, Republic of” I assume this is the same country? Should we add the 2 numbers together in the past months?
    • You still, as in past months, list “Russia” and “Russian Federation”, again this is one country, should we add them together?
    • It is not mentioned in any other month except February, as anything other than “Serbia” but in that month you have 2 listings, “Serbia” and “Serbia and Montenegro”. How should we handle that? Add them as being just “Serbia” or did the separate concern of “Montenegro” pop up in one month?

    • In Conclusion.

    I know this is long and detailed, sorry about that, sorry to dump a lot on you all at once. I have had most of these observations since you released the Key Metrics. This issue has always been important in regards to any meaningful research but it is in particular focus now with your new ‘what is a resident’ numbers and because I have to make reasonable predictions for my clients. Numbers I stand behind. I am designing and producing a large project for a Swiss client currently and it is essential for him and for me to make good plans based on solid, well defined numbers that have a vetted source and a well explained methodology of calculation. It has made all of us think about Second Life and the future of the metaverse with these numbers, it makes us realize that we are more pioneers than we thought. My client is in it for the long haul, as am I, he and I are not opposed to risk but risk is not the same as recklessness, we need to make predictions, plans and budgets.

    Like being a chef, statistical analysis is like preparing a recipe. You not only need excellent ingredients but you need to follow a set method and sequence of events or the meal is spoiled. I and others are willing to massage the numbers to squeeze more meaning out of them, make them tell their stories but we are not at the well like Linden Labs, we cannot see where the numbers come from in their pure form. For that we depend on you at Linden Labs. Doing this job well will be greatly appreciated and raise your stature and credibility. I have the highest confidence in your ability to carry out the task.

    Thanks, Loydin Tripp.

  54. Pingback: Dist

  55. Pingback: Romani activi in Second Life - suntem pe locul 23 « VIRTUAL INSANITY

  56. Pingback: Nordmenn brukte 53 000 timer i Second Life i mai

  57. From the SL Wiki:

    An Active User is defined as a Resident who has logged into the system for longer than 1 Hour for the reported time period.

    Is it correct to assume that the “active users” tab on the spreadsheet reflects the number of users that log in to SL every day? (I’m assuming as much, since the number of active users in a week is 20% or so higher, so this would make some sense). Is that an average, or the number of users taken on the day the metrics were compiled?

  58. Meta Linden says:

    @Loydin: good points. We do keep the links current for the downloadable files from our core website at the Economic Statistics Page. I have found, however, that by updating the reports on the core page and then blogging their update is a good way to have conversations about them – many folks use the RSS feed of the blog to get notified of updates rather than re-checking the website.
    As it so happens, I am working on a project to link over the metrics definitions on the wiki directly from the headers on the reports.
    the Country codes that we are using for the geography cleanup process are derived from the ISO Country code standard (3166-1, for those following along at home)- nothing against the CIA factbook data, we just lean towards the codes that the UN uses. 🙂
    @Gwyneth: The “Gender with Hours” tab shows the unique count of Residents (1.118M) and total sum hours spent inworld (20.767M) by all people who logged in during the month of May 2007. The Countries by Hours Tab and the Countries by Count and Age Bands tabs all show the unique count (507,844) and Total Sum Hours (18.195M) spent by Active users. Between these sums you can see that the “non-active” users- those spending less than an hour inworld over the month (I like to call these “browsers”) spent a total of only 2.5672M hrs, or around 12.3% of the time for more than half of the count of people. So none of these Active numbers are averaged or sampled, they are the sum counts.

  59. Pingback:   Tutti i dati ufficiali di Second Life by my Second Life

  60. Pingback: Pownce offers ‘transgender’ registration option; Will other sites follow? : gender equation

  61. Warda Kawabata says:

    Any chance of adding a column in the country pages, to show the number of avatars as a percentage of each country’s population (maybe display it as “active SL residents per 1000 people”)?

  62. Meta Linden says:

    Hi Warda-
    By releasing these as excel datasets, that’s exactly what I’m hoping people will do – connect these with country demographics, economic stats, broadband penetration rates, etc. In fact several bloggers are already doing that- see links to at least 4 of them on my blog. If any of you have another preferred demographic dataset you’d like to use for this, please post your findings too.

  63. U M says:

    How reliable are this stats, about spoken langugaes? How can you even measure them? The sence of measuring is silly and useless. Since anyone can get ahold of a translation device and or computer side appilation. Again how can anyone take these measurements serious? If they are based on the signup page how in the world can they be measured without major errors?

    Just a few thoughts here.

  64. Pingback: Barry Woudenberg » Blog Archive » # gebruikers Second Life overdreven

  65. Pingback: Networked_Performance — OFFSHORING SECOND LIFE

  66. Pingback: CSB / Confessions of a Second Life slacker

  67. Pingback: A framework for metaverses and SL. Part one. “Upstream: pros” « Fantasilandia

  68. Pingback: Movie Marketing, Improved » Blog Archive » From Salchow to Cash Cow

  69. Smsy says:

    Great work Meta. Great changes! Greatings

  70. elwis says:

    I agree with you Meta, thx for art

  71. Tv says:

    Good site and documans thanks site admin

  72. Tovya says:

    Meta, thank you very much for this information! I’m curious though if this demographic is updated to Q2 2008? The view of trends is astounding and I’m curious as to what happened middle of last year to make the Female #’s drop when it was showing nothing but an increase? I will keep searching for updated information from 2003 to Q2 2008!

  73. ikinci el says:

    this meta good thanks site admin wery nice

  74. sohbet says:

    Wery nice Meta thanks good works.

Comments are closed.