The traditional enterprise sales model is based on a well-oiled lead generation machinery that feeds a direct sales force. The formula is based on Total Revenue = Quota/salesrep * Number of salesreps. While this maybe an oversimplification, this is what it really boils down to. If you look at the Open Source business model, especially for an enterprise software company, the lead gen source are the downloads. The inside sales and eventually direct sales teams have to take these leads and convert them to sales. The big difference and reason why most of us are excited about Open Source is that for the first time emerging software companies are able to get a large volume of leads from successful Open Source communities. However, the conversion rates are low, perhaps as low as 1% over a reasonable period of time from download to deal close.
So, having said this, there is a play here if one has a very successful Open Source project that has thousands of downloads. If you don't have such a community, believe me, building one is not that easy. You are probably better off buying one if you can. Its an added risk especially for companies switching from a traditional enterprise sales model to an open source model.
For the more established players, this creates a new threat as now start-ups can start to spread virally (in a good way) in their installed base. Given the tens and thousands of customers they currently have, I would think that their strategy should be to emulate the open source model and offer low cost, low friction products and solutions that their customers can try and adopt. Once they are ready they can sign-up for support etc., the way one would do with an Open Source software company. On-Demand apps certainly address some of the issues being addressed by Open Source and a whole lot of issues that are not addressed by Open Source.
I hope this provides clarity to software execs looking for business reasons for adopting an Open Source strategy.
