Economy grows 94% year to year, hits new all time high in Q2
World expands and voice usage hits all time high; Bot policy affects user hours and logins

In Q2 2009, Second Life Residents enjoyed the sixth consecutive quarter of record growth and activity in the in-world economy.

At nearly USD 50M per month in user-to-user transactions, the Second Life economy is now on a annual run rate north of half a billion US dollars, making Second Life the largest virtual economy in the industry.  

The highlights of Q2 include:

• USD 144 million dollars in total user-to-user transactions, an increase of 94% from the same quarter last year

• Voice minutes totaled 3.2 billion minutes, up 48% from the same quarter last year

• LindeX and Xstreet hit new all time highs

• Resident-owned land in Second Life increased 11% from last year

Growth in total user hours for the quarter slowed, while monthly repeat users and concurrency declined, primarily because of a new Linden Lab policy limiting the use of automated, scripted avatars (Bots) used for the purposes of improving search results in Second Life. In contrast, economic indicators that track human activity (such as Voice minutes, LindeX transactions, and Xstreet gross sales) grew nicely.

Let's take a look at the data, trends and comparisons for Q2 2009, starting with the economy. 

User-to-User transactions increased 94% from Q2 2008 - The total of all transactions in Second Life reached a new high in the quarter.  The sum of all of the transactions in the Second Life economy equaled a total of USD 144 million dollars in Q2 2009. This is 20% growth over the previous quarter and 94% over the same quarter one year ago.

This means that in the space of one year, the Second Life economy has nearly doubled, which is a testament to the creativity, dynamism and energy of Second Life Residents.

By way of comparison, In Q2, the US Economy decreased 1% from the previous quarter and decreased 3.9% when compared to the same quarter a year ago (see page 40 of the full release and tables from the US Government's Bureau of Economic Analysis).

Trading Activity on the LindeX reaches a new all time high - The volume of exchange on the LindeX, the marketplace for Linden dollars (L$), the Second Life virtual currency, grew to a new all time high of USD 29 million in Q2. This represents 4% growth over the previous quarter and 8% growth over the same quarter a year ago.

Total Xstreet sales reach a new all time high in Q2 2009 - In its second quarter under the Linden Lab banner, Xstreet, the web marketplace for Second Life virtual goods, had gross sales of L$ 372M Linden dollars, or approximately USD 1.4 million.

This represents 13% growth over the previous quarter and 70% growth over the same quarter one year ago. Xstreet remains small by comparison with Second Life overall, at approximately 1% of the USD 144 million Second Life economy in Q2 2009.

User hours hit a new all time high of 126 million users hours in Q2 2009 - This is up 33% from Q2 2008 and up 2% over the previous quarter.  However, the monthly figures trended down slightly in the quarter, with April, May, and June clocking 43.2, 43 and 39.8 million user hours per month respectively.  This monthly trend is primarily the result of the new Bot policy introduced in late April. See the SPOTLIGHT section below for a detailed discussion of the impact of the new policy.

Second Life Residents used 3.14 billion voice minutes in Q2 2009 - Total Voice Minutes hit an all time high in Q2 2009, an increase of 3% compared to the previous quarter and 44% when compared to the same quarter a year ago.  This metric measures all VoIP traffic across all types of voice sessions - Resident-to-Resident calls, group voice chats, and local voice.

It is important to note that the bulk of Second Life voice minutes are Residents using local voice in small groups rather than person-to-person in-world "calls". As a result, measuring voice minutes in Second Life is more similar to measuring voice minutes for a teleconferencing service provider, because in teleconferences there are generally more listeners than there are speakers. However, as each user is connected to a voice channel, they are counted for the purpose of calculating total voice minutes.

Voice minutes grew faster year-to-year than overall user hours - 44% growth for voice compared to 33% for user hours. This is a function of two factors: increasing penetration of voice usage in Second Life; and the impact of Bots on user hours, which do not show up in the Voice minutes, as Bots do not use voice.

SPOTLIGHT: Bot policy change affects usage - Each quarter, we will drill down on a particular part of the Second Life economy that is timely or topical. We will rotate through topics and go deep into each area. Last quarter, the Spotlight section focused on land ownership and the effect of the Homestead/Open Space changes. This quarter, we will drill down on our understanding of the impact of our new policy regarding Bots.

For background on the change to our policies regarding the use of Bots to increase search ranking, please readconclusion to the blog post on Bots.  In short, the policy (effective April 23, 2009) made it a violation of our Terms of Service to use Bots to inflate traffic to inworld destinations in order to improve ranking in search. As a result, we have already observed what we believe is a change in the pattern of Bot usage.

In order to understand the impact of the new policy, we looked at the unique logins and user hours for June, dividing the Residents into usage bands based on the hours spent logged into Second Life each month. Here are the bands we chose for our analysis:

0 to 3 hours per month
4 to 50 hours per month
51 to 200 hours per month
201 to 300 hours per month
Greater than 300 hours per month

The below graph tells a very interesting story. When the count of users and the sum of total user hours by the amount of usage are calculated as percentages of their totals, you can see an interesting insight about Second Life usage:

There are very many light Second Life users who account for the minority of user hours, and the 18% of monthly users who spend more than 50 hours per month in Second Life drive 87% of the user hours.

This highlights the deep commitment and engagement of Second Life Residents, but it also raises the question of how much of that heavy usage is the result of Bots. This also points up the opportunity to increase the engagement of new users, a task which Linden Lab is focused on in 2009.

When you look at these numbers trended over the course of the quarter, it's clear that a change to small number of accounts with greater than 300 hours per could have a significant impact on user hours. A trended view for Q2 2009 makes the picture very clear.

The "greater than 300 hours per month" segment had a big impact on user hours in Q2. It is also interesting to note that the middle bands of 4-50 hours per month trended positive in the quarter.

To further understand the phenomenon, we took a sampling of 1182 accounts that with greater than 300 user hours per month were logged-in in May and were not logged-in in June, as these accounts were likely to represent accounts affected by our change in policy in April. Of these 1182 accounts, 86% had their last login using something other than an official Second Life release client, which is an indicator of an account being a Bot. Given that ~85% of all logins are with an official life viewer, we're relatively confident that the change in the usage pattern is due to our change in policy.

As a consequence, we expect that Q3 2009 user hours will likely be lower than Q2 2009 as we will have three months in the quarter - as opposed to just May and June in Q2 - without Bots. 

Turning to other usage metrics, we see a similar pattern, but the effect is not as pronounced as it is with user hours. 

Monthly Unique Residents with repeat logins peaks in May - In May 2009, Monthly repeat logins peaked in May at 752,035, and then declined slightly in June to 741,945. Year to year, June 2009 repeat logins grew 24% from June 2008 repeat logins. 

While the number of total unique logins (including users who log in to Second Life only once in a given month) is much higher, we use repeat logins (which captures all unique users with more than one login in a given month) as a metric to track the number of users who are engaged with Second Life.

Continued Stability supports the in-world economy - Our work on scalability and stability continued in Q2 2009 (seeUpdates from the Grid and Grid Monitoring and Support). In Q2, total downtime as a percentage of total user hours was .24%. This was slightly worse than Q1 2009 (at .21%), but on a par with the previous four quarters. Q2 also marked an 81% improvement over Q2 2008. The data for this metric does not extend any earlier than Q2 2007.

We measure total user hours lost to downtime (planned and unplanned) as a percentage of total user hours as a gauge of stability as well as a factor in the growth of the economy. Because this metric includes planned and unplanned downtime, it is key to understanding the total user hours that are available for economic activity in Second Life. 

We use this metric because a outage on Sunday at noon when concurrency is at its peak does not have the same impact on the economy as an outage Tuesday night at midnight when concurrency is at its lowest. For those Residents seeking a measure of system uptime independent of the total number of user hours or user hours lost to outages, the system uptime in Q2 was 99.89%.

Homestead sales drives addition of 109 million square meters - In Q2 2009 Resident-owned land increased to reach a total of 1.743 billion square meters.  This is 7% growth over Q1 2009 and 11% growth over the same quarter one year ago. 

The increase was driven primarily by the addition of new Homesteads in the month of June, which added approximately 98 million square meters to the world.   Many Residents purchased new Homesteads to take advantage of the $95/month pricing, which ended on July 1, 2009.

Using the CIA factbook as a source, Second Life is 189^th in the list of autonomous regions ranked by size, ahead of the Faroe Islands, Hong Kong and the Seychelles, and closing on Mauritius and the Virgin Islands.

Homesteads increase in June - Total Private Regions (Islands) owned by Residents reached 22,885 by the end of Q2 2009. This represents an increase of 8% over the previous quarter end in March of 2009. The chart shows that the bulk of the increase in private regions came primarily in June and primarily from the Homestead product.