If you sell them a Value Meal, they might come: A skeptical article in the San Jose Mercury News on whether the McDonald's Wi-Fi experiment in New York is a sign of things to come or a demonstration of the inefficiency of charging for hot spot service.
As regular readers know, I believe that captive venues like conference centers, airports, and hotels, have a high potential to get regular subscribers through networks aggregated for roaming. Non-captive venues, like public parks, cafes, or areas with retail density will certainly always have some amount of free service. But if you capture travelers with the promise of reliable service and they get the benefit of lower costs for Internet access in the captive locations, those same travelers will wind up using commercial service elsewhere if it's included in their monthly subscription fee.
That's to say that if you're already paying $30 to $50 per month for unlimited Wi-Fi because you spend a lot of time in Austin, San Jose, San Francisco, and Seattle, and hotel nights in Embassy, Wyndham, etc., you'll seek out the cafes that are part of the same roaming system because you'll be used to a guaranteed level of performance and customer service.