It's often difficult choosing what features to focus your attention on, and it's something I've definitely got wrong in the past. You react to the requests of a few, or you build what you think would be cool — but always run the risk of spending time and effort building something that no-one will use.
So is there a better way of choosing what's next?
Fortunately, yes - but it's not a quick solution. In essence; you wait, and you watch how customers use your product — then react to the patterns that emerge. Are people hacking your product to fit their needs? Are they doing something in 5 steps that could be achieved in 2? Can you see a way to make someone's life just that little bit easier?
If the answer is yes (and the answer is always yes) - then you can begin to create answers to genuine problems, using the patterns you've observed as your guard-rails for success.
That's not to say you shouldn't be constantly improving the small things as you see fit — but if your big features come from a genuine customer need, then they're far more likely to succeed.
Admittedly this is a method I've only just started using, and one that only really works when you're established enough to see the historical trends, and have the tools available to crunch the numbers. Fortunately for Prevue, with over 5 years of continued use, and a suite of admin tools — I can already see the benefit of taking the guesswork out of this process.