Survey

A survey is a structured set of questions sent to many people at once. Because it is cheap to run at scale, it can tell you how common something is: how many customers use a feature, how satisfaction differs between segments, which of five priorities matters most to the largest group.

What surveys are good at

Counting and comparing. If a persona assumption needs checking against several hundred real people, or you want to size a problem you have already identified, a survey is the right instrument. It is also useful for recruiting, since a short survey can feed a screener for later interviews.

What surveys are bad at

Be honest about this. A survey records what people say, not what they do, and the two differ more than most teams expect. People forget, flatter, round up and answer what they think you want to hear. It cannot tell you why, because you can only ask the questions you thought of in advance. And it cannot discover problems you did not anticipate, since there is no follow-up when something surprising comes up. Never use a survey to find out whether people can use your interface; watch them instead with usability testing.

In practice

Write questions about specific, recent behaviour (“when did you last do this?”) rather than hypotheticals (“would you use this?”), which flatter every idea. Avoid leading wording and double-barrelled questions. Pilot it on five people before sending. And pair it with qualitative work: interviews explain the numbers a survey produces.