EDIT: Joel's Camels and Rubber Duckies[1] is essential reading for how hard it is to sell enterprise software. Price it free, cheap, or dear. Not to say that there is no hope in selling to enterprises, because lots of people are very successful doing it, but you need to know how hard it is before you start.
[1] http://www.joelonsoftware.com/articles/CamelsandRubberDuckie...
We actually thought our lawyer was helping us but you know, he was a "close friend" with the CEO so the game rules were a bit faked.
Thanks for the book reference as well!
It also becomes a lot easier to sell your product to a big customer if you already have a lot of little customers. But selling software to big customers will require a lot of custom integration work: you need to have the scale to be able to provide integration services to them while still supporting your existing customers.
Enterprise customers should absolutely be a goal of any B2B software startup, but not early on. Early on you want to focus on the customers that are easier to please, and make a kick-ass product for them. Once you have built that product, you can start to scale it up to larger customers and get the feature integration right. You also shouldn't be afraid to say "no" if approached by a potential big customer because your product isn't ready for them.
Basically, there's a lot that can go wrong and you need to have a certain scale before you even attempt to go after the big whales. You shouldn't be trying to work with them while building out the basic version of your product.
I would often interact with a startup where the sales guy was fired on the other end because they just didn't understand that we were doing a 3-6 month competitive technical evaluation (in addition to our more important day-to-day work) after I'd clearly explained it a few times. No amount of "sales" will make up for a lack of required, costly-to-build features. Also, the answer to missing features is rarely ever "wait for them to be built."
An early-stage company really needs two things from its customers: cash flow and feedback. It's fine if you don't make any profits, but you need revenue coming in the door to get future rounds of funding. It's fine if your revenue per customer is lower than you would like. Small customers are great at both: price it low enough to start that it's not a huge business expense for them, and you will be able to deal directly with the people who will use the system. If they have a problem, your product guys will know and be able to relay it to development.
Big customers often get you neither. They require lengthy contract negotiations that will rack up thousands in lawyer's fees from your side before you bring in any revenue. They ask for features that aren't core to your product and are likely years down on your roadmap (think LDAP authentication, customer admin functionality, audit logs, etc.) Again, these have to be developed before they sign a contract and before you see money. If you actually manage to get them into a production environment, they're much less likely to give you actionable feedback.
tl;dr: B2B software startups that target big contracts are much more likely to fail because the market dynamics of chasing after big customers will increase your burn rate while diverting your sales team's attention from smaller fish.