Scott Rafer analyzes why free is cheaper to offer than for-fee hot spot service: This analysis is focused on the venue and operator, not the customer, in that billing and supporting customers costs, he says, $30/day/location versus $6/day/location for offering free service.

Scott's analysis is sound except for one glaring omission: free doesn't carry a promise of service with it, and the promise of availability, support, security, and quality is what will -- if anything does -- make commercial Wi-Fi hot spots a continuing reality. A free hot spot or network could disappear tomorrow based on the whim of the venue or operator and their changing model. Free doesn't involve a contract that works both ways: for value, we provide service. Free is one way: here you go, and take what you can get.

A for-fee system is more likely to be persistent and invest in innovation and special features because they have a motivation to retain and expand a subscriber base. But where does the subscriber base come from? From the kind of traveler who winds up spending a lot of time at captive venues: airports and hotels and conference centers. Captive venues may be unable to receive signals from Wi-Fi providers who aren't resident. And with reasonable 3G speeds further away (and some problems in using 2.5G inside concrete bunkers), captive Wi-Fi venues may remain captive.

As I've said many times before, these captive venues will always be able to decide whether to charge a fee because of the near-term lack of alternatives. Although John Yunker of Pyramid Research expects that most hotels will ultimately offer Wi-Fi and broadband for free as just another amenity within a few years, because it will be expected, like a dial tone, he misses a side issue, too.

Most analysts agree that there will continue to be a mix of fee and free hot spots, which means that business travelers will be forced to pay in many locations, especially airports. Which means that as prices drop, more travelers and corporations will find subscriptions more appealing, even if free service is available, because the free service may be in a forum non conveniens: a place hard to get to, as the lawyers call it.

Roaming is inevitable, and once you have a linked roaming network with predictable fees -- a la Sprint PCS's latest view of the world -- then you build subscribers who are willing to pay a fraction of their monthly cell phone bill for unlimited Wi-Fi. The price point has moved from $50+ per month down close to $20 per month with Boingo's recent price promotion and cell/Wi-Fi bundling deals by T-Mobile.

For-fee venues take on a different life when the cost isn't $6/hour or $10/day but $20 per month: even though the various venues get a smaller and smaller slice of that monthly fee, it still frees the venue from operating the network at their own expense.

We continue to focus on hot spots as a onesy-twosy service: travelers individually choosing to buy a connection or a subscription. iPass's IPO last week that gave them $100M in cash and $1B in valuation should reform our thinking. iPass resells access at a per-minute rate with some variation to entire corporations.

I do believe you could see locations that offer limited free service (email, restricted bandwidth adjusted dynamically, local location service, etc.) and expanded fee service (higher bandwidth, all service open, streaming media allowed) to subscribers of a network.

But as long as there are captive venues, a lack of broadband cellular data, an increase in roaming network agreements, and a widescale drop for unlimited service to the $20-odd/month level, we'll see a massive uptake in subscribers which will increase usage dramatically while lowering the venues' per-user take.

In effect, and ironically, many venues will probably see little enough revenue from for-fee services that it won't be part of their revenue calculations, but more focused on cost containment to retain passengers, guests, and members.