The first thing to do when designing a survey is figuring out exactly what you want to know. If you don't know the question, it's impossible to understand the answer. The second decision is your target audience. Who are you talking to? Asking the right questions of the wrong people will probably generate garbage.

The third decision is the methodology. Will it be self-completion online? In-person interview? Telephone interview? Each has its benefits and drawbacks.

Finally, you can start to put the questionnaire together. Creative Research Systems has an excellent tutorial on the creation and administration of surveys that even includes a link to a sample size calculator (this tells you how many people you need to interview to get specific levels of confidence).

Figuring out the questions to ask, and how to ask them, requires discipline. You have to be ruthlessly impartial, putting aside personal hopes and wishes, to generate neutrally-worded questions that will gain honest answers. It can be difficult, even painful, but that's the only way to get genuine information about your customers' (or potential customers') true needs. In a list of possible features, for example, you have to include ones you love, and ones you hate. You may be surprised what the respondents pick.

Finally, after all of the results are in, you have to look objectively at the data. That may hurt too, when you see that what you saw as the killer feature in your app has been greeting by a collective yawn from respondents. But that's the power of market research. You find out before you put in hours or weeks or months of sweat equity whether or not something will resonate with customers.