GigaOm rounds up the low subscriber numbers at early metro-scale Wi-Fi networks: They note the SF Chronicle story from yesterday about Taipei's underrun on necessary subscribers; point to Lompoc's 281 subscribers on a $3m network, where 4,000 are needed to break even (based on a 2003 analysis); and to criticism of MetroFi's first-phase Portland, Ore., rollout. EarthLink wouldn't give Katie Fehrenbacher usage numbers for their early networks.
Fehrenbacher concludes that mobile workforce applications will drive metro-scale networks--as Craig Settles noted--and that the way these early networks have been deployed, they don't provide either high speeds outdoors nor good indoor coverage.
As I've been writing for some months now, it's been very clear that the expectations set have been too high based on network density. It seems clear that to achieve indoor access rates at the level that residents expect, not the level at which a network can be financially and technically optimized, may be far above the density of nodes that are being built in most networks.
Of course, these are still early networks. Taipei is the largest single deployed network, although nothing like the largest network planned. They made some missteps in rollout, including having no compelling applications--not even a deployed VoIP service--that would change usage patterns. Lompoc has had equipment and vendor problems, and is rejiggering prices. Portland, Ore., is barely deployed, with just a few percentage points of the initial network installed.
St. Cloud, Flor.'s mayor just told me that every antenna in their network has had to be replaced due to water damage. They didn't pay for the antenna replacement, but city workers' time was apparently involved. St. Cloud scored highest in a recent network evaluation of several deployed U.S. Wi-Fi systems, but even they're having teething pains.
Update: The folks at Meraki sent me a link to the live overview of usage in their San Francisco network. As a free network with exposed stats, it's pretty interesting to see what usage is like on an ongoing basis.