Shopxare Team
The best customer satisfaction survey for an online store has two questions: one rating ("How satisfied were you with…?") and one open follow-up ("What is the main reason for your score?"). Send it right after the moment you want to measure: a support chat, a delivery, a return. Short surveys get finished: in Survicate's 2025 data, one-question surveys had a median completion rate of 100%, surveys with 4–6 questions 77.4%. Below are 25 questions to copy, sorted by moment, with the timing for each.
Pick the moment first, then the question
Most lists of survey questions give you 50 or 100 options and leave you to choose. The question follows from the moment. A rating sent after a chat measures the chat. The same rating sent two weeks after delivery measures the product, the parcel and the chat at once, and you can't tell which.
| Moment | What you measure | Best scale | When to send | Channel |
|---|---|---|---|---|
| A support conversation closes | That one conversation (CSAT) | 5 faces or 1–5 | Right away, 0–5 minutes | In the chat, or email for email tickets |
| Order confirmation page | The buying experience | Multiple choice | On the page, once | On-site |
| Parcel delivered | Delivery and packaging | 1–5 stars | 1–3 days after delivery | Email link |
| Return or refund done | How easy the return was | 1–5 | When the refund is paid | Email link |
| 7–14 days after delivery | The relationship (NPS) | 0–10 | Once per order, max every 90 days | Email link |
One survey per moment, and never two surveys for the same order in the same week.
25 customer satisfaction survey questions to copy
After a support chat or email (CSAT)
- How satisfied were you with the help you got today? (5 faces)
- Did we solve your question? (Yes / Partly / No)
- What could we have done better? (open, optional)
- How easy was it to get an answer? (1–5, very difficult to very easy)
- Was anything in our answer unclear? (open, optional)
Use 1 plus either 2 or 3. Five questions after a two-minute chat is too many.
On the order confirmation page
- What almost stopped you from ordering today? (Price / Shipping cost / Delivery time / Couldn't find information / Nothing)
- How did you hear about us? (multiple choice)
- Was there anything you couldn't find on our site? (open)
- How easy was checkout? (1–5)
Question 6 is the most useful question an online store can ask. Every "Couldn't find information" answer is a missing line on a product page or in your FAQ.
After delivery
- How satisfied are you with the delivery? (1–5 stars)
- Did your order arrive on the date we gave you? (Yes / No, it was late / No, it was early)
- Did everything arrive in good condition? (Yes / No)
- Does the product match the description and photos? (Yes / Mostly / No)
- What is the main reason for your score? (open)
After a return or refund
- How easy was it to return your order? (1–5)
- Why did you return it? (Size or fit / Not as described / Damaged / Changed my mind / Other)
- How satisfied are you with how quickly you got your money back? (1–5)
- Would you order from us again? (Yes / Maybe / No)
Question 16 overlaps with your return form. If you already collect a reason code there, drop it here.
Relationship (NPS)
- How likely are you to recommend us to a friend or colleague? (0–10)
- What is the main reason for your score? (open)
- What is one thing we should change? (open)
How to read the 0–10 answers is in our NPS guide.
Open follow-ups that get real answers
- What is the main reason for your score?
- What would have made this a 5?
- Is there anything else you want us to know?
- May we contact you about your answer? (Yes / No)
Ask 25 after low scores. A "yes" turns an angry rating into a conversation you can still fix.
Six wording mistakes and the fix
Pew Research Center, which writes surveys for a living, warns that questions asking about two things at once "are difficult for respondents to answer and often lead to responses that are difficult to interpret", and that agree/disagree statements pull people toward agreeing, which it calls acquiescence bias (Pew, Writing Survey Questions, 2021).
| Instead of | Ask | Why |
|---|---|---|
| How satisfied were you with our delivery and packaging? | Two questions, or only the one you'll act on | Two things in one question |
| Our support was fast. Agree / disagree? | How quickly did we answer? (Too slow / OK / Fast) | Agree/disagree pulls toward agreeing |
| How much did you love your new jacket? | How satisfied are you with the jacket? | Leading wording |
| Rate your experience (1–10) | How satisfied were you with today's chat? (1–5) | "Experience" of what? |
| Any feedback? | What would have made this a 5? | Vague questions get blank answers |
| All questions required | Only the rating required | Forced text gets "asdf" |
How long should a customer satisfaction survey be?
Two to three questions, under one minute. The data points the same way:
| Questions in the survey | Median completion rate |
|---|---|
| 1 | 100% |
| 2–3 | 86.8% |
| 4–6 | 77.4% |
| 7+ | 78.8% |
Source: Survicate, 2025 Benchmark Report, 8,392 website, in-app and widget surveys run in 2025.
The 7+ group finishing slightly better than 4–6 is a reminder that people who start long surveys are a self-selected group, not proof that longer is fine. SurveyMonkey's analysis of about 100,000 surveys found abandon rates rise once a survey takes more than 7–8 minutes, and that people tolerate less length when the survey is customer-related than when it is for work or school (SurveyMonkey).
Example: a pet supply store's after-chat survey
A pet supply store with three people on support sends a survey when a chat closes: question 1 (five faces) and question 23 ("What would have made this a 5?"), with only the faces required.
A customer asks: "Do you ship the 12 kg dry food bag to the islands?" The agent answers that it ships in 4–6 working days with a surcharge, and closes the chat. The survey card appears: three out of five faces. The comment: "Would have been nice to see the surcharge before asking."
That comment is not about the agent. It is a missing line on the product page. The fix takes ten minutes, and it is the kind of thing a longer survey with 12 rating scales would never have caught.
Survey emails and consent
A survey inside a chat the customer started is part of that conversation. An email survey is different:
- UK: the ICO says "genuine market research does not count as direct marketing", but if a survey "includes any promotional material or collects details to use in future marketing campaigns", the marketing rules apply (ICO, updated August 2025). Keep discount codes out of your survey email.
- Germany: the Federal Court of Justice ruled in 2018 (VI ZR 225/17) that a feedback request in an invoice email sent without consent counted as advertising and was unlawful (LTO report). Selling to German customers? Ask for consent at checkout or keep surveys inside the chat.
This is general information, not legal advice.
Setting it up in Shopxare
Shopxare has surveys built in (docs):
- Go to Surveys, pick Customer satisfaction (CSAT), NPS or Custom survey, and click Create.
- Under Questions, edit the text and set each Question type: Faces (1–5), Stars (1–5), NPS (0–10), Multiple choice or Open answer. Mark only the rating Required.
- For the after-chat survey, choose After a conversation is closed with a Delay (minutes) of 0. Messenger customers get a card in the conversation, email customers get an email.
- For delivery, return and NPS surveys, choose Only via a shareable link and put the link in your shop's delivery or refund email, or print it as a QR code.
- Set Ask the same person at most every (days), for example 30, so regular customers aren't asked after every chat.
- Click Activate. Nothing is sent before that.
Each survey shows Sent, Answered, Response rate, Average score and every answer, with Export CSV. The share of good ratings also appears as Customer satisfaction in Reports, next to first response time. How to read that number is in our CSAT guide. Shopxare is free, no credit card needed.
What to do with the answers
- Reply to every 1 or 2 within one working day, if the customer said you may contact them. Start with: "Thank you for telling us. Can you tell me what happened?"
- Tag comments once a week with 5–8 reasons (delivery, product info, price, agent, returns). The top reason is your next fix.
- Don't rank agents on CSAT alone. Many low ratings are about shipping or price.
When not to send a survey
- Right after you said no to a refund. You will measure the decision, not the service.
- To a customer with an open complaint. Fix it first.
- When you have nobody to read the answers. An unread survey is worse than none.
FAQ
What are the 5 key questions in a customer satisfaction survey?
For an online store: how satisfied the customer is (1–5), whether the problem was solved, how easy it was, the main reason for their score, and whether you may contact them. Don't ask all five at once. Pick one rating and one open question per moment.
How many questions should a customer satisfaction survey have?
Two or three. In Survicate's 2025 benchmark, 2–3 question surveys had a median completion rate of 86.8%, against 77.4% for 4–6 questions.
What is a good response rate for a customer satisfaction survey?
It depends on the channel. In-chat surveys sent right after the conversation get more answers than email surveys sent days later. Track your own rate per survey and compare it month to month instead of chasing a general benchmark.
Should I use a 1–5 or a 1–10 scale?
Use 1–5 (or five faces) for a single interaction like a chat or a delivery. Keep 0–10 for NPS, the "would you recommend us" question.
When should I send a customer satisfaction survey?
Right after the moment you want to measure: within minutes of a chat closing, 1–3 days after delivery, when a refund is paid. For NPS, 7–14 days after delivery.
