Writing Great Questions
Craft clear, unbiased questions that get honest, actionable answers
Your Progress
Section 3 of 5Every Word Matters
Writing survey questions is deceptively hard. One bad word biases the entire answer. "Don't you think..." leads the respondent. "Would you pay..." asks about hypothetical future behavior (people lie). "Rate your satisfaction and ease of use" asks two things at once (double-barreled). Good questions are clear, specific, and neutral.
Test Your Questions
Check Your Question
Bad vs Good Questions
"Don't you think this feature would be useful?"
Issue: Leading question
"How useful would this feature be to you?"
Why: Neutral phrasing, no suggested answer
"Would you pay $10/month for this?"
Issue: Hypothetical future behavior
"What did you pay for the last similar tool you bought?"
Why: Past behavior is more reliable than future intentions
"How satisfied are you with the speed and reliability?"
Issue: Double-barreled question
"How satisfied are you with the speed? (Then: How satisfied with reliability?)"
Why: One question per question
"Rate your experience"
Issue: Too vague
"How satisfied were you with the onboarding process?"
Why: Specific and clear
The 5 Rules of Good Questions
1. Be Specific
Bad: "How was your experience?"
Good: "How satisfied were you with the checkout process?"
2. Stay Neutral
Bad: "Don\'t you love this feature?"
Good: "How do you feel about this feature?"
3. One Thing at a Time
Bad: "How was the speed and reliability?"
Good: "How was the speed?" (separate question for reliability)
4. Ask About Past, Not Future
Bad: "Would you pay $10/month?"
Good: "How much did you pay for your last similar tool?"
5. Provide Clear Options
Bad: "Do you use it a lot?"
Good: "How often do you use it? Daily / Weekly / Monthly / Rarely / Never"
Rating Scale Best Practices
Likert Scale (Agreement)
Best for: Measuring agreement with statements
Example: "The product is easy to use"
Satisfaction Scale
Best for: Measuring satisfaction levels
Example: "How satisfied are you with customer support?"
Frequency Scale
Best for: Measuring how often something happens
Example: "How often do you use this feature?"
NPS (0-10)
Best for: Measuring likelihood to recommend
Example: "How likely are you to recommend us?"
Test Your Survey on 3 People First
Before sending to hundreds of users, test on 3 colleagues or friends. Watch them take it. Where do they pause? What questions confuse them? What do they interpret differently than you intended? Fix those issues before launch. 10 minutes of testing saves hours of bad data.
Key Takeaways
- •Be specific: "How satisfied with checkout?" beats "How was your experience?"
- •Stay neutral: Never use "Don\'t you think..." or other leading phrases.
- •One question at a time: Avoid double-barreled questions that ask about multiple things.
- •Test on 3 people before launching. Fix confusing questions early.