Case study of DarkWood themed community and RPC

DarkWood -- a fantasy/medieval themed community in Second Life


The challenge

A community with a recognizable style as its theme is taking a risk. Success or failure can be seen at a glance. DarkWood's fantasy/medieval theme has endured thanks to strong support by committed landowners, and Linden intervention during the early times. Now, however, Second Life has grown to a point where Linden staff intervention is becoming impractical. Will DarkWood stay true to its theme, or will the next big land sale open the door to malls or clubs like those on many main grid sims? This threat is real because DarkWood is so close to the busy Welcome Area with all those green dots on the map representing potential customers, the lifeblood of commercial establishments. A glance at the map makes this challenge apparent.


Background

In March 2003 Linden Lab selected five winners from the submissions received in response to a call for themed community proposals for new territories. DarkWood was one of the winners. The SL History Wiki (link) credits Nexus Nash with the original concept design. Tony 'Zero Grace' Walsh has posted an enjoyable article on his blog at this site.

He describes becoming a landowner in DarkWood, and quotes Nexus Nash's description of his original concept:

The original location of DarkWood was the Northern third of the Dore sim, which is one of the four Welcome Area sims. Since the DarkWood group had one of the winning proposals, LL gave group members land on a new adjacent sim to be called DarkWood. In the first year the SL role playing game called DarkLife originated on the new sim, and brought many people to DarkWood before finally moving to its own sim. Elf Clan's castle remains in DarkWood, serving as a recruitment center, and perhaps reminding DarkWood residents that the sim was more active before the highly successful Elf Clan moved to its set of private islands.


Governance

The officers of the DarkWood group are those residents who own land in DarkWood. They are expected to uphold the theme, and have had real power over land transfers. When an existing landowner wants to move out of the sim or tier down they receive the fixed price of L$1 per square meter for their land. The officers can vote on who is allowed to purchase this land for the same amound. In the SL Group Forums DarkWood is an anomaly since it has both an officers forum and a members forum to reflect the difference between landowning officers and everyone else. In a forum post David Cartier explains the political situation:

According to Rhiannon Chatnoir, one of the current guardians of the DarkWood legacy, the governance structures of this community include not only its set of written rules but also an unwritten and informal set of social norms. Apparently these informal practices work because the community has some continuity of long term land ownership, and values its heritage. This is unusual in a virtual world characterized by more population churn and the transience of neighborhoods.

I get the impression that declining Linden involvement has placed increasing importance on the informal norms. For example, Rhiannon mentions the original restriction on land sales within DarkWood as having been governed by an enforced covenant, but now '...a lot of the bigger structure isn't as solid... but still.. it is seen as horrific if you sell outside of the neighborhood without asking everyone.' She handles informal exchanges of IMs from people who hope she'll take care of abandoned tech junk on DarkWood land, and likens the process to a Neighborhood Watch.

Since DarkWood is coming up on its third anniversary it makes a good subject for long term observation to see whether any of the processes outlined in Ellickson's Order without Law (see thread in this forum) might be serving to bind this community.


Written Community Standards and Rules for DarkWood

Our community, when created, had several rules and regulations that not all members seem to be aware of. When it was formed the community was a small village completely ruled by officers and height restrictions were in place. Since the advent of 1.2 several changes have been made to better suit the community and also to preserve the theme. The Lindens in helping to establish this community granted the DarkWood and 1/3 of Dore to the community. This Group is a closed membership group, meaning that consideration of your presence in our membership, was thought out by one of its' members and saw you as a valued friend. If you are in our group, you are currently in one of the oldest, consecutively run group in SL since its' founding. The group has never disbanded or moved and we will always be pleased to have friends of DarkWood and will be ever greatful to all those who have ever helped in its' endeavors.

Now on to the rules:

1: Membership into DarkWood does not give you automatic land. A vote to permit land is called and after approval you will be notified of available locations until all of DarkWood is Purchased.

2: Our Community is a PG,Themed Fantasy/Medieval based society. Meaning that your style of building must be in a in this theme and all explicit behavior and conversations shoud be carried out in private or in the Mature sims. No Modern buildings of the Fantasy Genre.

3: NO Modern Weapons such as guns or Modern Vehicles, so we can preserve our communities Fantasy/Medieval style.

4: NO Building on Water that would cause said Waterways to be blocked unless approved by the Officers.

5: NO Land is to be sold to anyone outside of the community. It can only be sold to approved landholders and members and all land will purchases will be reviewed by the Officers. Land is bought and Sold at 1 Linden dollar per meter.

6: If you find you cannot hold land and no Officer is available, your land must be sold back to Nova or Haney Linden and notification to the Officers of DarkWood would be greatly appreciated.

7: The basic TOS is in effect in our community, such as it is in all and you should all ready be made aware of standards.

8: Any event of community theme, may be held and support will be given if needed. If you wish to hold an event on any of our properties, please contact the owner of that property and obtain their Permission. A considerable trust fund has been raised for this purpose. Contact the current DarkWood Liason on your need for funding.



Is DarkWood successful?

During a week of observation at different times each day I usually saw one or two additional green dots in the sim. There were no gatherings or events, just people exploring the builds and the marketplace.

Despite a history of occasional crises, including potential risk that land in DarkWood would be put up for sale on the open market, the community retains its fantasy/medieval theme. A small percentage of vendors in the marketplace display modern merchandise or signage, but there are no larger scale builds that are inconsistant with the theme. The most pleasant aspect of DarkWood is the prevalence of trees and gardens, including an underwater garden at one outer corner of the sim. People are allocating prims and space to beautifying the sim, and are not building McMansions up to the lot lines.

Comments are invited.

Your case study is pretty accurate as it stands. I just want to clarify some things. We have two groups because we were hoping to use the defective group voting function. We had so many members that putting up a proposal for a vote invariable failed to obtain a quorum, due to many members either never logging in or being so totally out of the information loop that they could be counted on to very militantly hold no opinion at all.
We formed the new group mainly to overcome that challenge, but even a small group voting on issues, proposals and the occasional recall was just as hopeless. We now make our big decisions by gathering and working out a solution that we all seem to think might work. So far we've been very lucky.
Likewise, the two forums were, as far as I can recollect, an accident of sorts, with both Soren and Zero requesting a forum and both getting one. We tend to use one for business and one for events, but since few actually read them, we don't use them much.
For a while, land transfers in DarkWood got completely out of hand. When one of our senior members was discovered to be not so very senior at all; he was a fourteen year old boy, in fact, he got booted by Linden Lab and his large tract of land wound up going to auction. We managed to keep the land in group but only at the expense of a shocking amount of money for the buyers.
Since the buyers paid cash for their land, rather than in Linden $$s, which I still tend to think of as so much Monopoly money, when and if they decide to pass on their stewardship of the land, they certainly won't be held to the covenant in pricing it for sale. I am not certain that they will ever get any return on their investment beyond being able to enjoy the land rather than having a mall in our midst. For me that is reason enough. I doubt they have any regrets.
Recently one of the major landholdings in the theme changed hands for the traditional 1L$ per meter, so hopefully we can hold onto that particular tradition that has allowed us to welcome so many new residents into our group.
Your study almost appears to paint us as a reclusive and exclusionary group. That really isn't the case - though some of us don't really get out much - as we have always welcomed visitors and actively recruited new residents, some of them fresh from Orientation Island. In a number of cases we've even subsidized their land purchases. To become a member of DarkWood is like becoming a Mason; you have to chase down a member and ask and you are in. During the time I have been a resident in DarkWood I can only think of two people we didn't let in and they were pretty notorious characters we were all familiar with from their actions in other regions.

Thanks for highlighting the problem of the group voting function, David. I knew that the larger group was unwieldy for decisionmaking due to members being hard to contact and so on, but you're making it clear that your group was actually trying to do votes with that tool. City Slickers, the Nova Albion residents group, uses it for messages, and I think Elf Clan does too. Of course, in my opinion, it's not very good for messages either. : )



My experience definitely supports your point. One thing stands out as anomalous: Elf Clan has a great facility with two signs on the front: 'ElvenWorks' and 'Join Elf Clan'. It's easy to see that there's a group looking for new members. However it's not so obvious that there's a group for DarkWood, particularly to new residents who aren't used to clicking on things to find the owner, looking at the group, and reading the charter of that group. Anyhow the charter text field doesn't say much compared to the notecard available from that notecard-giver sign on DarkWood castle. This isn't a criticism of DarkWood, of course. My own residents group in Nova Albion doesn't have a facility like that either. But maybe we should....

The material covering the history and characteristics of DarkWood is excellent. It would be interesting to have a little more detail on the economic aspect, if that would be possible. The original grant of land indicates that RL funds were not required to set this up - do the members have to pay tier on an ongoing basis? Also, have any commercial sites been added consistent with the overall theme?

David Cartier's comments about members having to buy land that was recovered by LL from an underage player seems to indicate there could be some circumstances where the members would have to come up with RL funds to maintain their distinctive environment. In terms of developing unique sims with less than mass-market appeal, the high cost of land is a major drawback. How much will a small group of people be willing to ante up for a hobby? A related topic might be the history of the five groups you mention that won the original LL contest. How many are still around, and comparatively, how successful have these groups been?

Much thanks to Jeska for enabling image attachments in our forum. Now that we have this capability I'm posting my DarkWood maps and screenshots to illustrate the following responses to Ferren's questions.



Being a member of the DarkWood community has no effect on tier. Apparently some members purchased lifetime accounts during the early days of SL, so they get free tier for 4096 square meters in recognition for their support during the crucial startup period of SL. Other members pay the usual rates.



DarkWood has extensive marketplace areas in the center of the sim. The marketplace areas were developed by private landowners, unlike the Linden-regulated marketplace facilities in Boardman, for example. Nearly all the marketplace signage and products are consistent with the fantasy/medieval theme. I noticed low key commercial activity outside the marketplace areas as well, for example a display of stained glass artwork also completely in harmony with the theme. Trees, gardens and parklike landscaping abound in this sim, including the marketplace area. Residents familiar with the appearance of typical main grid sims will appreciate how rare this is.



I cannot give a precise answer to this question, but a cursory inspection of the main map gives a clue about land values. You'll notice a substantial amount of yellow plots on the main grid, indicating land for sale. Some are silly priced attention-getting gimmicks, of course. You're unlikely to see any yellow plots in the DarkWood sim. Presumably people are not trying to unload their DarkWood land.



In SL the notion of 'successful' is subjective. From my biased perspective DarkWood is the most successful, because it's such an attractive leafy sim, and that's very hard to pull off when you have a number of independent landowners. As far as community spirit goes, my most recent experience, a St Patrick's Day celebration at Rhiannon Chatnoir's garden outlook, indicates that true friends of DarkWood live in many corners of SL, including the private islands, and not just the Elf Clan sims.

Financial success is an open question. The marketplace areas are never busy, but are they ever advertised? Rhiannon's event was the antithesis of the typical SL club event with typical advertising on the forums, and yet it was very popular without needing the expense of an event host or giveaways. In my opinion financial success might best be judged by some ballpark estimate of land values based on an inspection of the main map.

Here's an image of the old map:

Here's a recent image of the map:

Elf Clan has an effective recruiting facility in DarkWood. Note the outstanding texture.

Rhiannon's garden outlook with weathered stone ruins. Leafy greenery, natural landscaping and flowers make DarkWood far more desirable than the typical main grid sim.

This is true. There are currently four lifetime members involved with the theme. Not one of us is a Linden. Three of us have our 4096 tier allowances permanently donated to group land with other members. We have maybe 15000 square meters devoted to group land in the two sims and all of the land the three of us in DarkWood own is owned in common. This isn't such a hard move to make when you've known and trusted each other for nearly three years. A lot of land is also either owned outright or loosely affiliated with Elf Clan, as you pointed out. They take a similar route to a different destination but we all work very well together as a rule.




I recently sold my land in Dore to another returning DarkWood member for 1L$ for every square meter of land. As double prim land it is some of the most valuable land on the grid, but only if somebody is going to use it. My alternating prim-evil forests and continually unfinished builds were, though always in theme somewhat confusing and certainly not the highest and best use I thought possible for the land. The new owner, when he went away left a hole in our community and in our hearts and now he is home. That is how I value the land. None of us willingly sells our land in DarkWood. It is always the unavoidable crises in our first lives that force us to give it up.

David Cartier's comments about the continued involvement of original life-time members, and the benefit of their contributing their much greater tier towards the sim environment, tend to confirm my impression that the onerous nature of tier charges may be restricting the development of SL.

Unlike many other MMO's, where the appeal is to a mass market to attract everyone into the same troll-bashing milieu, SL has the potential to cater to smaller audiences. But those small audiences need to reach a critical mass to create sims, keep an active environment, and begin the slow process of adding other individuals with the same interests. An onerous, one might almost say punitive, tier structure will crush many of those initiatives soon after they begin.

From the number of sims purchased, kept in operation for a while, and then sold, I wonder whether LL isn't on the wrong side of the yield curve. Perhaps revenue would be even greater if more people kept more sims for longer. It would also make for an environment where these special interests could come together to build more lasting worlds.

Of course they are on the wrong side of the yield and it is short term thinking on their part. They are trying to recoup their initial investment in a server at once and then charge a great whacking fee each month on top of it. As a result there is a lot of turnover and increased administration work for Linden Lab. If they dropped the initial setup fee for a stock sim and then had the tier for an entire sim set at maybe half of what it is now there are an awful lot more people who could justify and then afford long term to buy and keep a sim for more than a month or two. The resulting revenue would be much greater than it is now. Maybe they just cannot keep up with that kind of demand now. We in DarkWood are discussing, have been for a while the prospect of a move to an island sim, much as Neualtenburg has done, but it is just out of our collective reach.

I wasn't around during the Neualtenburg transition to a private sim, but my understanding is that the project was fortunate enough to have members willing to lend it money to cover the startup costs. Those loans have since been repaid.

You said that taking Darkwood to an island was just out of reach. Had you considered the used private island market? You might have to wait a while to move the sim where you wanted it, but you might get a sim for less than the cost of a new one, Don't forget the owner transfer fee, though.

I would keep in mind that short lived groups are the rule on the Internet, and that long running cohesive groups, like Darkwood or the Gorean sims, are well out on the right hand side of the bell curve. (On reflection - IRL groups have a very similar curve.)


It's not an easy decision to price things such as full simulators ('private islands'). With an initial price and monthly fee too low, you end up with a great deal of capital tied up for months on end. This also makes your business vulnerable to churn - individuals or groups leasing a simulator for short periods, and then absconding. This leaves you with not only capital tied up - but also with a period of no income to compensate. A high rate on the other hand discourages this churn by making the purchase of a full simulator a 'serious' decision.

Given the bell curve effect noted above, this leaves the operator only two real choices - 1) deeply throttle the number of full simulators available for lease, or 2) charge an initial fee at or near your initial costs.

Elde Eponym wrote:



The economics of running SL would make an interesting study on their own, though as it is a private company we won't be getting access to the hard data.

We can roughly estimate some of the costs, and look at how they are affected by land, activity and avatars.

First is the cost of maintaining the application code. This is really independent of land or avatars, but is somewhat related to activity, as processing load and types of events stress the application and reveal program defects. There is also the initial development cost of the application, which is likely to take a number of years to recover. As the application does need to introduce some new features, there is also a ongoing component based on application development. Overall, we could look at this as a fixed cost that is amortized over the population of avatars and land. In other words, more participants means this becomes a lower and lower portion of the costs. The one part that would be driven by avatars is the "customer service" aspect of the operation.

Second is the physical server infrastructure to support the game. This is driven most directly by land, and secondarily by avatars. During the past three years, there has been a fairly significant increase in computing power in Wintel servers, and more importantly, new products like VMWare that allow hosting multiple Windows server images on one physical server. This allows much more efficient utilization of hardware. If the application is structured to allow multiple sims within a single Windows operating system image, there can be additional savings by reducing the effect of the Microsoft Tax.

Thirdly, there is the required network infrastructure to allow the online world. This will be driven directly by activity, and secondarily by number of avatars.

If I were designing a pricing structure around this cost base, I would look at recovering the application and network costs primarily over the user fees paid. The land sale and tier should be sufficient to cover the server infrastructure required. This involves purchase of initial server capacity (which could be from $500 to $800 per sim using the latest equipment and software (10 server images per $12K to $15K server), and if the application could support three or four sims in one server image). The infrastructure support costs per sim would be on the order of $50 to $100.

One limiting factor here would be the existing investment in individual PC and Windows licenses. That has become an inefficient way to provide the environment, but there is likely some amortization period before replacement can occur. The land charges have to include a provision for periodic replacement of hardware and software, and you can't do that with one time fees.

Another way to look at this is to ask someone with actual project management business experience in costing out a growing common carrier server farm running user-modified applications to include core infrastructure install and maintenance with hundreds of servers, thousands of concurrent users, administrative, software development, marketing and return on capital loads.

After all, my recollection is that Mitch Kapor and the other LL investors are not in this for fraternal, religious or charitable purposes. They are trying to build a business that will return their capital with profit sufficient to cover their investment risk.

Any one here (other than LL) have that experience?

Frank Lardner wrote:



Some of what I do in RL is reviewing cost models for these kind of services (I'm not a finance type, so I don't claim to understand all the NPV and IRR bells and whistles). After running through the rough numbers in my post, it struck me that there probably isn't a lot of room to lower the tier price on private sims given the current environment.

And yet, and yet... You look at Guild Wars, one of the most popular games going now, and they must operate a huge server farm with literally millions of players and vast amounts of content, all for... free. Pay $39 for the game and you can play GW for years. The one factor that would have massive impact on the financial model would be the number of environments that can run within one Windows image, or indeed whether you even need to use Windows and pay the Micro$oft Tax.

If LL wanted to make dramatic reductions in cost, it would probably be best to:

1. convert the environment from Windows to Linux;
2. either upgrade the server farm to virtual servers, or revise the code to allow multiple sims to run on one operating system image;
3. bring in some top-notch application tuning experts to go through the code and make it run faster.

The latter is often overlooked, so long as the code runs. However, I've seen applications that suddenly go down to using a few percent of the former CPU cycles where some programming inefficiences are cleaned up.

I believe the servers are LAMP boxes, so LL has already done that. I believe the equation is not one sim per server, but one sim per processor. Someone correct me if I'm wrong.