Customer Satisfaction Survey: Questions, Templates and AI Guide

The key word is structured. Casual feedback — a support email, a passing comment, an app store review — is valuable but unsystematic: it over-represents the angriest and happiest customers and cannot be compared over time. A survey asks every customer the same questions on the same scale, which turns scattered opinions into a measurable signal.

Businesses run satisfaction surveys to answer practical questions: Are we meeting expectations? Where does the experience break down? Which problems are most common? What should we improve first? Because responses are quantified, the answers become evidence you can act on — and a baseline you can track after every change you ship.

Satisfaction is not a vanity metric — it is an early-warning system for your revenue. Dissatisfied customers rarely complain before leaving; most simply stop buying. A regular satisfaction survey surfaces those risks while there is still time to respond.

• Retention and churn — satisfaction problems show up in surveys before they show up in cancellation numbers. Falling scores in a segment are a leading indicator of churn.
• Loyalty and repeat purchases — customers who consistently report good experiences are the ones who come back, upgrade and renew.
• Referrals and reputation — satisfied customers recommend you to others; dissatisfied ones tell even more people. Surveys help you find both groups early.
• Product improvements — open-text answers tell you which fixes and features customers actually care about, in their own words.
• Support quality — satisfaction data by channel and by issue type shows exactly where service breaks down.
• Revenue risk — a structured survey lets you quantify how much of your customer base is at risk, instead of guessing.

Timing determines both response rate and answer quality. Ask too early and the customer has nothing to evaluate; ask too late and the memory has faded. Match the survey to the moment:

Transactional moments (purchase, delivery, support) call for immediate surveys about that specific experience. Relationship moments (milestones, renewals) call for periodic surveys about the overall relationship — see the difference explained in our customer satisfaction vs NPS guide.

A good survey mixes a small number of question types deliberately. Each has a job:

• CSAT rating — “How satisfied are you with your recent experience?” on a 1–5 scale. The standard headline metric for a specific interaction.
• Likert scale — “The product was easy to set up” from Strongly disagree to Strongly agree. Good for testing specific statements.
• Multiple choice — “What were you trying to accomplish?” Gives you clean, comparable categories for segmentation.
• Yes/no — “Was your issue fully resolved?” Unambiguous checks on a single fact.
• Open-ended — “What is the main reason for your rating?” The single most valuable follow-up in any satisfaction survey.
• NPS — “How likely are you to recommend us?” on 0–10. Measures overall loyalty rather than one interaction.
• Customer Effort Score (CES) — “How easy was it to get the help you needed?” Effort strongly predicts frustration and churn.
• Quality ratings — “How would you rate the quality of the product?” Focused on the deliverable itself.
• Support experience — “How satisfied were you with the support you received?” Scoped to the service moment.
• Loyalty and repurchase — “How likely are you to purchase from us again?” A forward-looking check on retention intent.

For a deeper breakdown of each metric and when to use it, see customer satisfaction metrics: CSAT, NPS, CES and more.

Below is a compact question bank organized by purpose. Pick a handful that match your goal — never all of them. Short surveys get better data.

This nine-question structure works for almost any business as a general satisfaction survey. It runs about two minutes and every question earns its place:

1. 1How satisfied are you with your overall experience? (1–5 scale, required)
2. 2How well did we meet your expectations? (1–5 scale)
3. 3How satisfied are you with the quality of the product or service? (1–5 scale)
4. 4How easy was it to accomplish what you wanted? (1–5 scale)
5. 5How satisfied were you with any support or service you received? (1–5 scale, with a “didn't contact support” option)
6. 6How likely are you to purchase from us again? (1–5 scale)
7. 7How likely are you to recommend us? (0–10 scale)
8. 8What is the one thing we should improve? (open text)
9. 9May we follow up with you about your feedback? (yes/no, optional)

The core structure stays the same across industries — what changes is the trigger, the wording and the insight you track:

Full example surveys for each industry — with triggers, questions and matching templates like the restaurant experience survey, guest satisfaction survey and patient satisfaction survey — are in our industry examples guide.

No single number captures satisfaction. Mature programs combine a small set of complementary measurements:

• CSAT — the share of respondents choosing the top ratings for a specific experience. Compute yours with the free CSAT calculator.
• NPS — loyalty on the 0–10 recommendation scale (NPS calculator).
• CES — how easy it was to get something done (CES calculator).
• Average rating and distribution — a 4.0 average made of 4s tells a different story than one made of 5s and 1s. Always look at the spread.
• Response rate — low rates raise the risk that respondents don't represent your customer base (response rate calculator).
• Qualitative themes — recurring topics in open-text answers, the layer that explains every score.
• Sentiment — the emotional tone of written feedback, useful for triage at volume.
• Issue frequency — how often each specific problem is mentioned, which drives prioritization.
• Satisfaction by segment — scores split by plan, region, channel or customer age reveal problems that averages hide.
• Trends over time — the direction of your own scores matters more than any external benchmark.

The full workflow — from choosing metrics to closing the loop — is in how to measure customer satisfaction, and the metric-by-metric reference is in the customer satisfaction metrics guide. If you're deciding between the two headline metrics, read customer satisfaction vs NPS and customer satisfaction vs CSAT.

• Define one clear objective per survey — a survey that tries to learn everything learns nothing.
• Keep it focused: 5–10 questions for transactional surveys, and say how long it takes upfront.
• Ask the most important question first — completion drops with every screen.
• Use neutral wording: “How would you rate…” not “How much did you love…”.
• Avoid double-barrelled questions — “Was the product easy to use and good value?” is two questions.
• Keep scales consistent across questions and across time, or your trends become meaningless.
• Include at least one open-ended question — it is where the explanations live.
• Send at the right moment for the experience you're asking about (see the timing table above).
• Optimize for mobile — a large share of responses will come from phones.
• Tell customers why their feedback matters and what you did with the last round.
• Never ask for information you already have — it signals you don't know your customer.
• Test the survey yourself, on a phone, before sending it to anyone.
• Segment results by customer type, channel and lifecycle stage before drawing conclusions.
• Close the feedback loop: respond to dissatisfied customers individually where possible.
• Act on recurring themes and track whether scores move after each change.

• Asking too many questions — every additional question costs completions and data quality.
• Leading respondents toward the answer you want to hear.
• Using confusing or inconsistent scales (1–5 in one question, 1–10 in the next).
• Combining two topics in one question, which makes answers uninterpretable.
• Sending the survey too late, after the memory of the experience has faded.
• Surveying the same customers too frequently — fatigue destroys response rates.
• Collecting feedback and never acting on it — customers notice, and stop responding.
• Treating NPS, CSAT and CES as interchangeable — they measure different things.
• Ignoring open-text feedback because it is harder to process than scores (AI fixes this — see below).
• Reviewing only the average score and missing a polarized distribution.
• Failing to segment — an acceptable overall score can hide a segment in crisis.
• Shipping a survey that breaks on mobile.
• Asking sensitive questions (income, health, personal data) without genuine necessity.

The two hardest parts of running a satisfaction program — writing unbiased questions and analyzing hundreds of open-text answers — are exactly what AI is good at.

• Generate a complete survey from a goal — describe what you want to learn and get a structured draft in seconds, adapted to your industry.
• Improve unclear or biased questions — AI rewrites leading and double-barrelled questions into neutral ones.
• Create logical flows — follow-up questions that appear only when relevant, so happy customers finish fast.
• Summarize open-text responses — every written answer read and condensed into themes.
• Identify recurring topics and their frequency, so you know what to fix first.
• Analyze sentiment across responses to triage at volume.
• Surface pain points and recommend concrete next actions grounded in the actual responses.

On Surveys.expert, this is the built-in workflow: describe your goal in AI Studio, review and edit the generated survey, publish it, and let the AI analysis summarize themes and sentiment as responses arrive. The full walkthrough is in our AI survey builder guide, and the analysis side is covered in how to analyze open-text responses.

1. 1Define what you want to learn — one objective, written down.
2. 2Choose the trigger: which moment in the customer journey is this survey about?
3. 3Select your primary metric — CSAT for a specific experience, NPS for the relationship, CES for effort.
4. 4Add three to six focused follow-up questions from the question bank above.
5. 5Include one open-ended question that asks for the reason behind the rating.
6. 6Customize the design so the survey looks like it comes from you.
7. 7Test it yourself on desktop and mobile before sending.
8. 8Publish through the right channel — email after transactions, in-app for feature feedback, link or QR code elsewhere.
9. 9Analyze both the numbers and the open-text answers (or let AI do the reading).
10. 10Turn findings into actions, tell customers what changed, and measure again.