How to Write Survey Questions Users Can Actually Answer

The most dangerous survey questions are the ones that get answered anyway. Users will politely respond to questions they do not understand, about experiences they barely remember, and the results will look like data. Surveys That Work by Caroline Jarrett is the antidote: a craft guide to questions that produce truth.

What makes a question unanswerable?

Three common sins. First, asking about things users cannot know: “How likely are you to recommend us next year?” Future intent is fiction. Second, double-barreled questions: “Was checkout fast and easy?” Fast and confusing gets which answer? Third, jargon and abstraction: “Rate your satisfaction with our self-service ecosystem.” Users guess, you analyze the guesses, decisions follow.

What are the wording rules that work?

• One idea per question. If it contains “and,” split it.

• Anchor in recent behavior. “Thinking of your last order” beats “in general.” Memory distorts fast.

• Use the user’s words. If customers call it a “top-up,” do not ask about “balance replenishment.”

• Label every scale point. “Satisfied” to “dissatisfied” with named points outperforms a bare 1 to 5, because everyone interprets the points the same way.

Why do answer options matter as much as questions?

Because users choose from what you offer. A question about missing features with the wrong options teaches you nothing except what you already assumed. Jarrett’s advice: build options from what real users said in interviews, not from what the team brainstormed.

Why does Tremoloo share this book?

Because question craft is where survey quality lives, and it is the part teams skip. We use Jarrett’s standards in every survey we design for clients, and the difference in data quality is immediate.

Want surveys that produce decisions, not dashboards? Our research team can write and run them.