Surveys That Work: Why Most UX Surveys Fail Before They Launch

Every product team has a drawer full of survey results nobody used. Caroline Jarrett’s Surveys That Work explains why: the failure happens before launch, at the moment someone opens a survey tool and starts typing questions.

What is the core lesson of the book?

A survey is not a list of questions. It is a decision-making instrument. Jarrett’s method starts where most teams finish: what decision will this survey inform, and who will make it? If nobody can answer that, the survey should not exist. This single filter kills most surveys at conception, which is exactly the point. A survey that cannot change a decision is not free research. It is a cost imposed on your users and your team’s credibility.

The book’s other great insight is the total survey error perspective: your data quality is only as good as the weakest link among sampling, questions, and response. Teams obsess over question wording while surveying whoever happened to open the app that day, then treat the results as truth.

What do designers take away from it?

• No decision, no survey. Write the decision first, in a sentence.

• Users can only answer what they know. Satisfaction with a feature they used once last month is not data. It is noise with decimals.

• Pilot every survey. Questions that seem obvious fail in surprising ways. Test on five people before five thousand.

• Acting matters more than asking. A survey you will not act on teaches users that their feedback disappears.

Why does Tremoloo share this book?

Because surveys are the most abused research method we encounter: easy to launch, hard to do well, and disastrous when trusted blindly. This book is the discipline we wish every team applied before pressing send.

Want research that informs decisions instead of filling dashboards? Our research team can design it with you.