I have done a version of this. When a client has negotiated a payment schedule after work was completed, I set up "passwords" to arm the program between payments. When a payent was made the next passward was issued until all monies were recevied. If payments were missed then the program expired. This worked OK but I didn't like it.

This idea is certainly worth pursuing. My concerns would be:

How about enhancements/modifications asked/demanded by the client? Are you obligated to do enhancements/modifications for no increase in the monthly rent?
Will your support costs be covered? i.e Is $50/month/site enough?
Is there a minimum period the client must pay this fee to cover your dev costs?
Is ther a maximum period after which the client will not have to pay?
Are you to cover training their employees?
Who owns the app?
Can you license this app to your client's competitors?
What are you liabilities if 24/7 service is not avaiable? Or not available during biz hours?
Several dozen locations times $50 /month could be a nice steady income but supporting several dozen locations is also a steady expense stream? Are expenses really covered with enough extra for you to host all of us at DevCon?

The major risk I see with the little info I have is you are fronting the development cost with only a promise of gettingyour money back from the client. They may pull the plug before yu go "on-line" or stop after a few months. The major reward I see is the client loves it and pays the several dozen monthly fees for all eternity.

It is an interesting concept.

Just a few concerns.

>Hey folks,
>
>I may soon be working on a relatively simple piece of software that will be installed in several dozen locations. The interesting part of this deal is that the client is unwilling to pay an upfront per-site fee but is willing to pay $50-75 per month per location license and support fee.
>
>Each of the locations has a DSL connection. What I have thought about doing is to have the application validate the license once a month (or when the paid-through dat is approaching) by "pinging" a license server that would maintain paid-to-date status per client. On installation, the site would receive a unique ID and the new installation info would be sent out to the license server (more than likely) as an XML packet.
>
>I'd like to see some comments about this approach: A) Is it technically feasible to do reasonably and, B), does this sound like a solid model for deploying and marketing software?