Today, Linden Lab is (finally) publishing its much-anticipated Q2 metrics for Second Life, something I have been pushing for over the last few months. Last night, I transcribed the material the Lab provided, and added observations and commentary to it. You’ll find that below.

For the lazy, the rushed, and the hard-of-thinking, here’s a quick summary:

  • Reporting has changed. Linden Lab isn’t publishing all the old familiar figures, particularly a couple that took a swan-dive during the quarter. This is because they are ‘not indicative’, apparently.
  • User-to-user transactions are deprecated (again). Apparently the Lab has trouble figuring out which figures are and aren’t ‘indicative’.
  • Certain statistical feeds are to be shut down. They will be ones that the Lab feels are not ‘indicative’.
  • Linden Lab indicates that the economy is stable, but there aren’t sufficient metrics provided to support that position.

Everything else requires more explanation, so on with the show.

Caution: Contains numbers, statistics, metrics and numerical indicators Much of this text is as the Lab gave it to me yesterday, and Linden Lab is also a source of the graphs presented here (although I had to hand copy the text, so it is possible I have introduced a typographical error somewhere). All of the Lab’s text will be marked in italics. The remainder is my own commentary.

You may click on any of the images to see a larger version, rather than squinting at the scrunched-down inline images.

Second Life Economy Stable in Q2

The Second Life economy remained stable in the second quarter of 2010, despite one-time events that affected a number of metrics. The most notable economy event of the second quarter was a temporary decrease in the exchange rate of the L$ vs. the US dollar in mid-June that has since returned to typical levels. Our research in the weeks since this occurred indicates this change was due in large part to our merge of Residents’ inworld and XStreetSL L$ balances. This had the temporary effect of reducing L$ purchase volume as Residents lowered the total amount of L$’s they were holding because they now had one combined account. Previously, Residents held a higher combined L$ balance in two separate accounts and purchased L$’s more frequently when they were keeping a minimum balance in each.

Firstly, I’ll jump in and say that I’m pleased to see that the Lab isn’t reporting massive economic growth where there demonstrably isn’t any.

However – and it is a mighty big ‘however’ – this quarter’s metrics report follows a pattern usually used by corporations to obfuscate or conceal bad results. Therefore, it seems logical to me to assume that bad results are being obfuscated and/or concealed.

I’m seeing what looks like smoke, and I wonder where the fire is, exactly.

Changes in Reporting

From time to time we will change our reporting to provide the most relevant and accurate summary information for the use of merchants, landowners, content creators and other participants in the Second Life economy. Our goal is to provide those who participate in the economy and run inworld businesses with the data they need to make informed business decisions. We are making a change from the recent past by reducing the relatively large set of metrics we have provided – many where causality is complex or unclear – and we will instead provide a more focused data set with a briefer description about each quarter’s activity. We believe the indicators below are the key items to focus on and we are adding a page describing the content and methodology behind each figure.

This has come up before, on more than one occasion. The idea that the numbers are confusing and don’t clearly support the message, so we’ll stop showing those numbers to you and show you some other numbers instead. Yes, this sort of thing makes me nervous.

Also, yes, the numbers can indeed be rather easy to misinterpret. Linden Lab has itself done so on a number of occasions. Not the least reason for misinterpretation being a long-held lack of transparency about how the numbers are gathered and processed.

In addition, we will be phasing out the data feeds at http://secondlife.com/statistics/economy-data.php. We’ll have more updates on this shortly, but because this data currently bears little relationship to the health or strength of the Second Life economy, we feel it’s appropriate to discontinue these feeds.

Frankly, I’m not astonished at this. This is probably the least surprising surprise in today’s report.

Resident-to-Resident Transactions

At least for the time being, one figure we will no longer publish in our historical charts is the Resident-to-Resident transaction data. As we’ve analyzed this data we believe that it is not as accurate a measure of the inworld economy as we would like. We are currently working to develop a more accurate metric with which to replace this number. For the near term, we will continue to publish a comparable Resident-to-Resident transaction figure for historical comparison. In Q2 2010, Resident-to-Resident transactions totalled approximately L$40 billion, or US$152 million.

You heard it right. User-to-user transactions are being deprecated by the Lab as a metric for economic activity… again. In previous years, the Lab had taken great pains to thoroughly discredit the metric as a valid economic measure, and then pulled a 180 later on. Now it seems the Lab has performed ‘a 360’ and is shooting down that figure again.

This doesn’t seem like a bad move, but not replacing it with a new metric this quarter seems problematic.

RESIDENT USAGE INDICATORS

MONTHLY REPEAT LOGINS

Monthly Repeat Logins indicated improved retention and engagement over previous quarters:

  • Q2 2010 averaged 805 million monthly repeat logins, a rise of 8.6 percent vs. the Q2 2009 average of 742 million.
  • Levels are flat with Q1 2010.

Repeat logins started to fall immediately after 31 March this year. Could that be attributable to the release of Viewer 2? I can’t think of any good reason.

If you’ll observe the graph, you’ll see that “flat with Q1 2010” basically means reversing much of the gains made in that quarter.

USER HOURS

User Hours were slightly lower, at 108 million for the quarter.

  • The Q2 2009 User Hours figure of 126 million includes substantial usage attributable to “bots” and other scripted agents. Bot policy changes took place in mid-Q2 2009, making it difficult to make accurate comparisons with the year-ago period.
  • User Hours in Q1 totaled 116 million.

It’s now been a year since the Bot policy changes. Someone uncharitable might think that the fact that it keeps getting brought up means that the Lab has been highly ineffective at implementing the policy, and thus bot-adjustments continue to roll on quarter after quarter.

If, on the other hand, the Lab has been effective in dealing with bots, then there’s something else wrong here, because there is a distinct decline.

Actually, it’s likely that there is some other cause than just bots in the decline of user-hours, but the text given invites you to think that bots are the only significant cause.

MONTHLY ECONOMIC PARTICIPANTS

Monthly Economic Participants (those accounts that receive or distribute L$ in a given month) are the most active and engaged participants in Second Life and the economy. Economic participation in the second quarter, on average, was flat against the previous quarter, and rose against the same period a year ago.

LINDEN DOLLAR RATES AND RELATED METRICS

EXCHANGE RATE

The average Linden dollar exchange rate moved from approximately L$261 per US dollar at the beginning of the quarter to just under L$269 per US$ at the end and fluctuated across a larger range for two weeks during June (see second chart below), as we noted at the time. Average exchange rates, however, remain largely unchanged. As noted above, our research indicates this change was due to our merge of Residents’ inworld and XStreet SL L$ balances and temporary market uncertainty immediately after the announcements of our restructuring and changes in senior management.

L$ SUPPLY

We are also improving and changing how we report the L$ Supply figure with this quarter’s post. Previously, this metric did not include individual XStreet SL account balances, and misclassified some administrative accounts as Resident accounts. For this quarter we have updated the calculation to include only Resident accounts, and have summed their balances across both XStreet SL and inworld. This number now more accurately represents the total purchasing power of all Residents in Second Life. Note that we have shown both calculation methods in the charts.

  • L$ Supply in Q2 was US$78.2 million vs US$77.9 in Q1

You’d think that represents an uptick, wouldn’t you? The graph that Linden lab supplies tells a different story, showing a reduction in L$ supply in Q2 vs Q1, and quite different figures indeed.

If we are to believe the graph, L$ Supply in Q2 was US$25.9 million vs US$26.2 in Q1. So the text gives figures that are 200% larger than the graph, and are going in the other direction.

LINDEX VOLUME

Historically, LindeX volume has been very stable and the slight decline from Q1 is within past ranges. This metric is traditionally less volatile than many other indicators we report and we expect this behavior to continue.

  • LindeX trading volume of US$29.4 million in Q2 2010 was up slightly against the year-ago Q2 volume of US$29.1 million, but off from Q1’s US$30.7 million

It’s notable that the figures for the LindeX do not include sales of fresh L$ from Supply Linden, an important stream of revenue for Linden Lab and is arguably the best indicative metric for Second Life economic health that have traditionally been available. (There may be better figures that the Lab has never published)

According to the Lab, Supply Linden only creates new Linden Dollars to satisfy certain buy orders on a demand basis. If the economy keeps growing, demand keeps generating new Linden Dollars to match that growth, and Linden Lab makes (or made) millions of US Dollars from that.

That particular data feed has been completely pulled. I believe the Supply Linden sales figure for Q2 is zero (that’s where it was headed in Q2, and there’s no data now to contradict that trend), which indicates an economy in decline, rather than an economy which is stable or growing.

XSTREET SL MERCHANDISE SALES VOLUME

Q2 2010 saw 39% growth in sales over Q1 2010 and 124% growth over the same quarter a year earlier. The primary driver of this growth was the merging of users’ Xstreet SL and Second Life Linden dollar balances.

Oops. Someone at the Lab forgot to change ‘users’ to ‘Residents’ there for the final draft.

This simplified the purchase process on Xstreet, producing an immediate jump in sales volume. In addition to the benefits of the balance merge, we continue to show healthy growth from the acquisition of new users paired with the development of new creative products from our merchants. Q2 continued the trend of setting record sales volume and this quarter in particular contained an important milestone, as Xstreet SL had it’s first month with over US$1 million worth of sales.

This could well dovetail with declining user-hours, as you could reasonably argue that time spent shopping on XStreet SL is not time spent engaged in-world.

LAND

WORLD SIZE

World Size grew to 2,110 square kilometers[sic] at the end of Q2, from 2,070 at the end of the previous quarter, and 1,900 for the year-ago period. Private regions accounted for the largest portion of inworld land (44 percent), and private homesteads the next largest (36 percent). Mainland regions accounted for 19 percent of the total inworld land mass.

The percentage of sims that constitute mainland has remained remarkably consistent for several years now (hovering between 19 and 20%). That’s something I frequently wonder about: whether mainland growth is managed so that it scales as approximately to non-mainland growth or whether the consistency in results is largely incidental.

During Q1 2010, the average price of land per square metre actually sold was L$0.87/sqm.

During Q2 this fell to L$0.71/sqm (and is presently about L$0.59/sqm).


Overall, the impression the Lab gives is that the Second Life economy is in decline. Not a huge decline but a decline nevertheless. Omitting several metrics make it hard to justify the position that the economy was largely stable during Q2 2010.

Top of my list of metrics I’d want to see added are:

  • The median price of open mainland sold per square metre.
  • Supply Linden’s sales figures (along with a ratio of aggregate sources and sinks).
  • Median user session-length per day.

I’d also like some enormous bags of money, a holiday, and the mortgage paid off. But I don’t think I’m going to be seeing those in a hurry either.



Got a news tip or a press-release? Send it to [email protected].
Read previous post:
Close