The value of dirt in Tokyo

Tokyo Skyline, seen from Tokyo TowerI keep running across people trying to make the case that land in Second Life is overpriced, because it costs more than the corresponding proportion of the cost of setting up a disconnected OpenSim instance on an Amazon EC2 virtual server and paying someone to answer the phone, or whatever.

And that’s just silly.

You can’t compute the “natural” cost of a piece of land in Tokyo, or Manhattan, or London or Paris, by working out how much it would cost to buy the corresponding number of truckloads of dirt on the open market.

If the economics were really as these simplistic analyses want to claim they are, then we would actually be seeing a large migration of people off of Second Life and onto the “cheaper but just as good” alternate grids that were profitably springing up everywhere.

And we aren’t.

Yeah, there’s lots of interest in other grids, hundreds of people are checking out InWorldz and Reaction Grid and so on, and setting up their own OpenSim instances on Amazon EC2 servers, but these are actually alternatives or competitors to SL in only a tiny number of cases. Mostly they are things that people do in addition to SL, or they are things that people do who wouldn’t be in SL anyway.

Like a piece of land in Tokyo or Manhattan, Second Life isn’t about the dirt, it’s about the location (and the location, and the location, to bow to the legally required truism). Linden Lab does not host generic servers, or generic regions, or even generic customer support. It hosts Second Life, and Second Life is something that lots of people are willing to pay (in time, in eyeballs, and even in US$ or equivalents) to take part in. And that, presumably, is because it has people and events and builds and experience and wonders and riches and connotations and associations that aren’t to be found in those cheaper generic places.

You could actually do a relatively simple calculation of the location-value of SL, by taking the cost of an SL region, and subtracting that “generic OpenSim region running on Amazon EC2 plus 1% of one generic offshore support person” number that people love to drag out.

But instead of framing that number as “see look SL is overpriced by $X!”, you’d have to say “wow, people value SL over the alternatives by $X!”. And somehow that’s not the conclusion that the people usually doing this math seem to want to reach. ;)