Some questions related to the selling of Enterprise Software..
Mathaix - the agreement you described is fairly common, because it does make sense for customers. If you're a business, and you've identified that there isn't a product out there that meets your needs, you're faced with the decision of working with a non-ideal product, or convincing someone to make what you want. If you can help someone else start a business around your idea, you've incentivized them to build a tailored solution for you, and then be around to support it. If you were to just build it yourself, you'd be responsible for ongoing improvements, additions, etc. This would just be a distraction since the product is likely not your core business.
For giving away the software, that's what pilot projects are for. In an enterprise sale, the custom expects the roll-out to be pricey, so it makes sense to minimze the risk. For example, if you're doing a $1MM software project, it makes sense to do a $50k pilot project to learn some lessons and increase the liklihood of success. Let them know you're giving the software for free (or better, only charge them for 1-5 user licenses, and let them know it'll count as a discount towards the final project budget), but charge for the time. If they aren't willing to pay for the consulting/pilot work, then they aren't likely going to invest in your software down the road.
Re generating leads, network, network, network. You, as the founders, are best suited to sell the fist few clients. In fact, you need to sell the first few (you're likely the only people that can). The expensive sales person w/ rolodex is only once you've proven the sales process, the model, and that you can support the clients properly. Go to trade shows, join associations, do web-based seminars, etc. But remember, you'll need to make the first sales yourself.
This is one of the best ways to get started as an Enterprise Software Company if you're not able (or don't want to) raise capital to begin with. Just be careful about two things: 1) be really tight on the IP ownership and 2) don't build non-standard stuff. Make sure the software comports to a strategy that you want to pursue. Many companies just build it because they're asked to rather than knowing it's a market opportunity they want to pursue.
I blog about enterprise sales and entrepreneurship in general (as well as VC). My series on enterprise sales in the link
Easiest way to sell is make your product accessible and easy to get started with. I work for an open source enterprise software company and never before have I been involved with a product that was so easy to sell into large enterprises. We have thousands of community users which have the potential to become customers. Our marketing team can track downloads and hits on our documentation for leads. Most customers that buy are already in a pilot or possibly production phase and hungry to do a deal with us!
just make sure your enterprise offering is diiferentiated and adds compelling value over your open sorce one or you will be competing (and losing) with yourself.
To address your question #1, there are a few reasons companies will do it. In some cases, they just don't care who else owns it because they're not in the business of selling software and it's not a key differentiator. They might also realize that someone else will be paying to make the software better so as long as they are on maintenance or get free upgrades, it's in their best interest for you to have multiple customers. If you're a small company, they may see it as insurance to help you stay in business.
Question #2 is key - if IT doesn't have the budget in terms of dollars or time, it doesn't matter if you are giving it away. The business side can sometimes push hard enough to overcome this, but it's a matter of priorities, so if your product is a nice-to-have rather than a have-to-have, it will be difficult. You can mitigate this by making the installation process as simple and self-contained as possible, maybe via an SAAS approach or a software appliance or VM. There may still be issues related to security, i.e. critical data living outside the firewall. There may also be a magic number at which a manager can expense a purchase rather than have to go through IT or a budgeting process; sometimes you can sell a product for under this amount with a lot less hassle - at least to get in the door.
As far as leads, it's already been stated that networking is key. It will be difficult to sign up an enterprise customer from a cold call. Find a customer you can partner with on the first one that will be a reference for future sales.