Startup founder Tim Romeo decided to kill his startup right before they would have gotten a check for $500,000. Sounds crazy? No; he did the right thing:
[S]omething was wrong. It seemed trivial at first, but it bothered me. Despite glowing praise, our users were only using ContractBeast to create a small percentage of their total new contracts.
I spent the next two weeks visiting our beta users, looking over their shoulders as they worked, and listening to them explain how they planned on using the product. Pressing them directly on why they were not using ContractBeast to create all their contracts resulted in a lot of feature requests.
Now, talking with customers about features is tricky. Often you receive solid and useful ideas. Occasionally a customer will provide an insight that will change the way you look at your product. But most of the time, customers don’t really want the the features they are asking for. At least not very badly.
When users are unhappy but can’t explain exactly why, they often express that dissatisfaction as a series of tangential, trivial feature requests. ... These aren’t necessarily bad ideas, but they had nothing to do with why they were not using ContractBeast more extensively.
His blog post is good advice not just for startup founders, but for anyone writing software.