Otherwise, just talk to people at parties. Everyone works with some horrible system.
People talk about the "information revolution" as if it's over. Look around at the amount of paper forms still being pushed around -- it's hardly even started.
Look at GitHub, who presumably have giant enterprise accounts by now. But they started by making something great for hackers. While enterprises might seem like cold places devoid of humanity, a lot of people inside the enterprise want to get stuff done as much as anyone else. They will pay out of their own pocket and claim the cost as an expense later. They will fight internally to adopt products, etc. And once an app is used by a team, if it's good, it has the potential to spread virally inside the company.
Other examples: Evernote (who just launched Evernote Business), Google Docs (the word processor came from Writely), Yammer, Atlassian's Confluence. I expect Trello is already used all over the enterprise and will be turning profits on that too.
Don't believe me? Good. Treat them as hypotheses and test them. You will be astonished.
(Except for the last bit about why businesses buy; you're right there. The reason that big company decision-makers buy is not that the new product is "technically" better. They buy because they are convinced that the new product will deliver a better return on investment and also convinced that the decision will not blow in their face and get them fired.)
I guess I must have misunderstood, because this is not worthwhile advice :)
How would you test this? Yes, one way is to build B2B products without any input and then try to sell them. Many people try to do this. This is, as you pointed out, a really terrible approach.
A much better approach is just to hit the streets or the phones and talk to potential customers. Maybe you think they'll blow you off (and in my experience, yes, some of them will). But what I'm asking you to test experimentally is that 100% of them will blow you off. I have actually tried this, and the surprising result is that this is simply not the case.
But don't believe me. Try it for yourself.
Once you have people who will talk to you, then you can find out their problems. Once you know what their problems are, you can see if other people in that position have similar problems. Pick the most painful one of these problems, and, with your new industry insider friends to guide you, build an awesome product.
My big point here is that one should not simply accept limiting beliefs about what is and is not possible without actually putting them to the test. Do the uncomfortable thing. Grow.
They would want to change because "somewhat works" is exactly that--somewhat works. The software that I mentioned in the post above took 45 minutes to export a single site report. The reporting was so bad that they actually had accepted that they would need 6 full-time positions in accounting just to reconcile the reporting and make progress sheets for their 500+ investors.
They are desperate for an alternative solution but they are already dealing with the least-worst software available.
The best way to sell products is to sell your time to the company, assess it's problems, and then help to solve them with software you've written.
Per 9000 HN posts to this date, you don't want to be someone who writes software, you want to be someone who solves problems.
Customers want their problem solved in a specific way, they don't care about the OS or language of the month.
Talk to people. Friends. Family. People who have jobs. Who can intro you to their managers. Ask questions. Listen listen listen. Spend your time out of the office, out on the streets, doing customer visits. Read and learn everything about an area you might want to solve. Succeed.
In enterprise B2B, nobody's going to give the time of the day to a startup whizzkid just because his solution is technologically better or faster or whatever, without connections and financial backing (stability).
Enterprise is more about who knows who. Painful sales cycles. Talking to loads of people across many business layers (while greasing the right pockets, dinners, gifts and whatnot).
Anonymous B2C is popular with startups for a reason.
The scenario then is this: is it easier to achieve the massive scale needed to make a B2C startup profitable, or is it easier to spend time breaking down the iron curtain?
Seems to me the latter problem is solved with domain expertise, which you can hire. There is no silver-bullet to the former problem.
Your last point is valid but misses one point I was trying to make. The low-hanging fruit in the enterprise world is not solving problems that have already been solved shittily (like creating a competitor to LotusNotes, which is almost universally hated). I agree, that's a tough sell.
The low-hanging fruit is solving problems that haven't actually been solved yet. Sure, you need to spend time in middle or back office at a big company to see these problems but so what? Go out, get a job, discover problems.
You can't expect to dream up solutions out of thin air.