Today I stumbled across Phil Wainewright's "How to get return BEFORE investment" post from 2005, and am surprised by how aptly it describes not only Software as a Service (SaaS), Wainewright's subject, but also open source:
With on-demand applications, customers don't start paying until they begin using the application, and they typically pay on a per-user per-month basis. So it's quite easy to imagine deploying a procurement application, for example, which achieves enough savings per user in the first month to more than repay the monthly fee. If the fee is billed on net 30 day terms, then the customer achieves the return before the investment has even been made. That's the essence of rBi [Return before investment].
In the case of open source, it's very likely (indeed, probable) that an enterprise will start paying for support or add-on functionality before going into full production, as the development stage requires the most hand-holding by the vendor.
But the principle remains: an enterprise can make full use of the software before paying a dime to the vendor, thereby de-risking the IT investment. No need to pay until you're relatively comfortable with the software in question.
This is the new enterprise software model, whether it manifests itself as open source or SaaS. I suspect that it will roil many an incumbent enterprise software vendor, as they struggle to make the economics pay for them, as CIO.com recently noted in its review of ERP vendors' SaaS strategies. but ultimately customers and vendors will benefit from a switch to subscription-based models that keep vendors honest and customers happy.