Don't listen to users literally. Establish metrics and track that.
Most people are too nice to tell you, "No, I am not going to use your shitty product". Instead, they will say, "Oh, that's cool. Can you give me feature X, Y and Z as well?". Then, you spend few weeks developing that. Again, they will then ask for X1, Y1 and Z1. It's endless cycle.
2 months into it, i saw the patterns and started asking. "Okay, if i give you X2, Y2 and Z2, will you use it?". That's when they said "No" and i had an even more fundamental problem.
Lesson: solve real problem.
There is a great documentary movie that shows that raise and fall of a company in the 2000 crash. The movie is called startup.com
http://www.imdb.com/title/tt0256408/ The movie shows several mistakes that you can learn from, including the need for a good founders agreement, the need for good security in the office, what can happen if you are over optimistic when dealing with VCs and several other lessons. I don't want to spoil the movie, so I won't explain the lessons farther in here.
I addition, there are HBS cases that talk about failures. One such case is Chemdex.com ( http://cb.hbsp.harvard.edu/cb/web/product_detail.seam?R=898076-PDF-ENG&conversationId=693223&E=55956 )
From a personal failure, I can recommend that you don't waste money on patents until you get an investor to pay for them. Patents are not worth much if the are not for a chemical or drug, and they cost tons of money. If you believe that you really need one, file a provisional patent application, and maybe file for PCT2 status as well. That is rather cheap (a few grand) and it gives you 30 months to decide if you want to file requests around the world or not. During the 30 months, look for an investor, and if you cannot find one, drop the idea. While there is a one in a million shot that your patent will be worth money some day even if you cannot find an investor now, but winning the lottery with get you better returns on the investment.