How to Write Good Survey Questions (With Examples)
Write neutral, specific survey questions that get honest answers — with before-and-after examples of the common mistakes.
A survey is only as good as its questions. Sloppy wording doesn't just lose you answers — it produces confidently wrong data, because people answer the question you accidentally asked. These are the rules that separate questions you can act on from numbers that only look like insight.
Ask one thing at a time
The most common mistake is the double-barreled question — bundling two things into one. The respondent who feels differently about each half is forced to pick a side, and you can never tell which half drove the answer.
| Don't ask | Ask instead |
|---|---|
| How satisfied are you with our price and support? | Split into two: one question on price, one on support. |
| Was the checkout fast and easy? | "How would you rate the checkout speed?" + "…how easy was checkout?" |
Keep the wording neutral
A leading question telegraphs the answer you want. So does a loaded one, which smuggles in an assumption. Strip out the adjectives and the flattery and let people tell you what they actually think.
- Leading: “How much did you enjoy our award-winning onboarding?” → “How would you rate the onboarding?”
- Loaded: “What did you like most about the new dashboard?” assumes they liked something → ask whether they liked it first, then why.
- Absolutes: avoid “always” and “never” in the question — they push people toward the safe middle.
Be specific and time-bound
“Are you happy with our service?” is too vague to act on. Anchor the question to a concrete experience and a timeframe, so every respondent is answering about the same thing.
- Name the thing: "the support reply you received," not "support" in general.
- Bound the time: "in the last 30 days," not "ever."
- Avoid jargon and internal product names a customer wouldn’t recognise.
Match the answer options to the question
Half of a good question is good options. For closed questions, the choices need to be mutually exclusive (no overlap) and collectively exhaustive (everyone fits) — and a rating scale should be balanced, with as many positive points as negative ones.
Keep it short, and order it well
Completion drops with every question. Ask only what you will actually act on, lead with the questions that matter most (in case people bail early), and save optional or sensitive questions for the end. Grouping questions onto separate pages makes a longer survey feel manageable instead of like one endless scroll.
Frequently asked questions
What is a leading question?
A leading question nudges the respondent toward a particular answer through its wording — for example, "How great was our fast support?" assumes the support was great and fast. Neutral phrasing ("How would you rate our support?") lets people answer honestly.
What is a double-barreled question?
A double-barreled question asks about two things at once, like "How satisfied are you with our price and delivery?" A respondent who loves the price but hates the delivery can’t answer truthfully. Split it into two separate questions.
How many questions should a survey have?
Fewer than you think. Every extra question lowers completion. Ask only what you’ll act on, group related questions onto pages, and put the most important questions first in case people drop off.
Should I make questions required?
Require only the questions you genuinely need answered; make the rest optional. Forcing an answer to a question someone can’t honestly answer produces noise, and too many required fields increases abandonment.
Related resources
Put these questions into a real survey
Build a multi-page survey with choice, rating, and text questions, add a description to each, and share it by link or embed. No backend, free to start.