Customer Feedback Survey Template (Free)
Launch a short, touchpoint-specific customer feedback survey and get a score you can track over time. Pick one core metric (NPS for relationship, CSAT for a just-finished interaction, or CES for task effort), add 1-2 fixable drivers, then review results weekly and follow up with low scorers who opt in.
Trusted by 5000+ Brands
Pick the right core metric: NPS vs CSAT vs CES
| Metric | Best fit touchpoint | Standard question (keep stable) | Typical scale | How to interpret | Do next |
|---|---|---|---|---|---|
| NPS | Relationship / loyalty after a meaningful period of use (quarterly, post-renewal, after onboarding completion). | "How likely are you to recommend [Company/Product] to a friend or colleague?" (see standard NPS wording guidance) | 0-10 (Promoters 9-10, Passives 7-8, Detractors 0-6) | NPS = % Promoters - % Detractors. Use it to track internal trends and compare segments; avoid treating any single number as universal. NPS is popular, but research shows the relationship to growth depends on context, so focus on your own trend line and drivers. | Pick 1-2 drivers tied to the moment (value for money, product reliability, support quality), then ask one open-text "why?" Follow up with detractors only if they opt in. |
| CSAT | Interaction satisfaction right after a completed event (ticket closed, call ended, delivery received). | "Overall, how satisfied were you with [the support you received / your delivery / your purchase]?" | 1-5 (or 1-7) agreement/satisfaction scale. Keep your choice consistent; see Likert scale options (5-point vs 7-point). | Higher is better. Track % satisfied (top-2 box) or average score. Read it as a pulse on that event, then drill into what caused low scores. | Add 1-2 fixable drivers for that channel (wait time, resolution, communication), then route low scores to a queue for review. |
| CES | Task friction for a specific job-to-be-done (checkout, onboarding step, returns, password reset, self-serve support). | "How easy was it to complete [task]?" (or "[Task] was easy to complete.") | 1-5 or 1-7 ease/effort scale (define endpoints clearly) | Higher ease (or lower effort) is better. When CES drops, fix the step that shows up most often in comments (forms, confusing UI, handoffs, missing info). | Add one driver that points to the step (clarity of instructions, number of steps, time to complete) plus one open-text prompt for friction details. |
Default setup: use one primary metric per short survey (NPS or CSAT or CES). If you need more detail, add 1-2 driver questions and one open-text "what should we change?" prompt instead of stacking multiple top-line metrics.
Next step: pick the single moment you are surveying (relationship vs interaction vs task), then lock the core question text and scale so your results stay comparable over time.
Evidence-based notes (useful for keeping your program consistent)
- Use metrics for internal trending first: Customer metrics move when timing, audience, and channel change. Keep your method stable and prioritize within-program comparisons over time (see ISO 10004 customer satisfaction monitoring guidelines).
- NPS is context-dependent: Treat it as a trend-and-driver tool, not a universal benchmark (see Keiningham et al. (2007) on Net Promoter and revenue growth).
Customer feedback questions you can copy (by touchpoint)
Use this section as a modular menu: pick one core metric module (NPS or CSAT or CES), then add 1-2 drivers that match the touchpoint, one open-ended prompt, and (optionally) follow-up consent + light segmentation.
Quick module map (choose what matches the moment)
- Relationship / loyalty: NPS + value/price driver + open-text (run on a consistent cadence like quarterly or post-renewal).
- Post-support: CSAT + resolution + speed/effort driver + open-text (send right after a ticket/call/chat ends).
- Post-purchase / delivery: CSAT + on-time + order accuracy + open-text (send after delivery is confirmed).
- Task or flow (checkout/onboarding/returns/self-serve): CES + step-level clarity/friction driver + open-text (keep the task wording specific).
- Renewal / cancellation: NPS or CSAT + value-for-price driver + open-text (use the same definition of "active customer" every run).
"How likely are you to recommend [Company/Product] to a friend or colleague?"
Why it matters: This is your relationship/loyalty read. Use it when you want a steady trend line across months or quarters, not a score for a single ticket.
When to use: Relationship surveys (post-onboarding completion, post-renewal, quarterly). If you only have room for one question, pair this with the open-text "why?" prompt below.
"Overall, how satisfied were you with the support you received today?"
Why it matters: CSAT tells you if a just-finished interaction met expectations. It is ideal for channel and agent coaching because the event is fresh.
When to use: Transactional surveys right after ticket closure, call end, or chat end.
"How easy was it to complete your purchase today?"
Why it matters: CES is your task-friction detector. When scores fall, you can usually fix a specific step (forms, payment errors, unclear shipping costs).
When to use: Checkout, onboarding steps, returns, self-serve flows. Keep the task description specific so feedback points to a fix.
Keep wording single-idea and non-leading so you can trust trends and comparisons. If you are rewriting items, use survey question best practices as your checklist (one idea per question, plain language, defined timeframe).
"Was your issue resolved?"
Why it matters: Resolution is a driver you can act on quickly. Low CSAT with low resolution points to process gaps, not just tone.
When to use: Post-support (email/call/chat). Pair with CSAT or CES, not all three.
"The time it took to get help was acceptable."
Why it matters: Speed is a common, fixable driver. If this is low, you can adjust staffing, routing, or deflection without changing your core metric.
When to use: Support moments. Use when wait time is a known complaint or a key KPI.
"My order arrived when I expected it to."
Why it matters: Delivery timing is a high-volume driver for ecommerce and retail. It separates "late" from "poor communication" when paired with a comms driver.
When to use: Send after the delivery event is confirmed (not at ship time).
"The product I received matched what I ordered."
Why it matters: Accuracy issues drive returns and contacts. This question points to pick/pack and catalog problems you can fix with ops work.
When to use: Post-delivery and post-pickup surveys.
"I reached my first success milestone quickly."
Why it matters: Onboarding success is often about time-to-value. Low scores usually mean unclear setup steps, missing training, or a product gap.
When to use: After your defined "first value" event (not day 1). Write the milestone in your internal doc so you can keep it consistent.
"The product is worth the price I pay."
Why it matters: This is a renewal driver. It separates product value issues from support issues when your NPS/CSAT moves.
When to use: Pre-renewal or post-renewal, and in cancellation flows as a driver check.
"What is the one thing we should change to improve your experience?"
Why it matters: This turns a score into an action list. You will get repeated themes you can bucket and assign owners to.
When to use: Include in every run. If you need a second open-text item, add it only when you have time to tag and act on comments.
"May we contact you to follow up on your feedback?"
Why it matters: An explicit follow-up opt-in lets you close the loop without surprising customers. It also reduces risk if you store contact details.
When to use: Use when you have a real workflow (owner, SLA, tracking) to respond to low scorers.
"Which of these best describes you?"
Why it matters: Light segmentation helps you find where the experience breaks (new vs long-time customers, plan tier, region) without overloading the survey.
When to use: Optional at the end. Keep it short (1 item) unless you need deeper targeting.
Customize and deploy without breaking your trend line
- Pick one primary metric and lock it: Choose NPS (relationship), CSAT (interaction), or CES (task). Keep the core question text, scale, and scoring the same across runs so your trend line stays comparable.
- Change one major element at a time: If you must change something, change only one of these per version: wording, scale, timing, or audience. Version your survey name (v1, v2) so you can explain any break in the trend.
- Send based on the touchpoint event: Internal starter targets (adjust after your baseline): post-support within 15-60 minutes of ticket closure; post-purchase 1-7 days after confirmed delivery; onboarding right after your first value milestone (not at signup). If you change timing, treat it as a trend break and re-baseline.
- Set a frequency cap to prevent fatigue: Starter default-adjust after your baseline and contact patterns: no more than 1 invite per customer per 30 days for transactional surveys (support, delivery). If customers have frequent interactions, consider a longer cap such as 60-90 days to reduce over-surveying.
- Keep it mobile-first: Internal starter target: 6-10 questions max, with one open-text item (adjust after your baseline completion rate). Put the open-text prompt after the core metric so you still get a score when people drop off on mobile.
- Keep segmentation optional and minimal: Internal starter target: ask 0-2 customer attributes at the end (plan tier, region, product line), and only if you will use them in reporting (adjust after your baseline drop-off rate).
- Use consistent invites to reduce bias: Keep invitation channel, sender name, and subject line steady when you are trending results. Starter default-adjust after your baseline: send one reminder max for transactional surveys. Track nonresponse bias and response bias when response patterns shift.
- Sample from the same event definition every time: Invite everyone who hits the same trigger (ticket closed, delivery confirmed) or a consistent random sample of that group. If you only survey edge cases, your scores will drift for reasons unrelated to experience quality.
For practical fieldwork rules (clear invitations, consistent procedures, and careful handling of missing responses), align your setup to a standards body such as AAPOR's Best Practices for Survey Research. If your response rate drops or your respondent mix changes, treat it as a data-quality issue first and adjust sampling and contact strategy (see National Academies guidance on nonresponse).
How to analyze results and turn feedback into action
- Score your core metric the same way every time
NPS: % (9-10) minus % (0-6). CSAT/CES: track average score and a simple "top box" rate (for example, % 4-5 on a 1-5 scale).
- Starter default-adjust after your baseline: Build a one-page trend view showing the last 8-12 weeks (transactional) or the last 4 quarters (relationship).
- If you change wording, scale, timing, or sampling, note it on the chart as a methodology change.
- Create score bands you can act on
Starter default-adjust after your baseline and segment sizes: Use 3 bands so routing and analysis stay simple.
- Low: NPS 0-6 or CSAT/CES 1-2
- Middle: NPS 7-8 or CSAT/CES 3
- High: NPS 9-10 or CSAT/CES 4-5
- Segment by touchpoint and channel to find the break
Start with the segment you can change: channel (email vs chat), queue/team, product area, carrier, device type, or onboarding path.
- Look for the biggest, most persistent gaps between segments (example: one channel consistently scores higher than another), then read those comments first.
- Only compare segments that have enough responses to be stable in your program.
- Tag open-text into reason buckets
Starter default-adjust after your baseline comment volume: Use 5-10 reason buckets and tag each comment with 1-2 buckets.
- Example buckets: "slow delivery," "confusing UI," "agent helpfulness," "billing issue," "missing feature."
- Starter default-adjust after your baseline: Keep bucket names stable for at least one quarter so you can see whether fixes reduce specific problems.
- Prioritize fixes by impact x volume
Impact = how strongly a bucket concentrates among low scorers vs high scorers. Volume = how often it appears. Fix the highest-impact, highest-volume bucket first.
- Starter default-adjust after your baseline capacity: Commit to 1-3 fixes you can ship in the next 2-4 weeks.
- Avoid a backlog of "insights" with no owner, due date, or measurement plan.
- Close the loop weekly (and only contact people who opt in)
Close-the-loop works best when it is operational, not occasional.
- Starter default-adjust after your baseline: Run a weekly 30-minute review, assign an owner and due date for each top issue, and track status in a simple action log.
- Starter default-adjust after your baseline: Follow up with low scorers who opted in within 1-3 business days, document outcomes (resolved, goodwill, refund, coaching), and re-measure after changes to confirm your internal trend moves.
Who should take this survey (and who should not)
Send this survey to current customers right after a specific experience you can name and trigger: ticket closed, delivery confirmed, onboarding milestone reached, renewal completed, or cancellation submitted.
Do not blast your full customer list unless your goal is a relationship check-in (NPS-style) and you can define who counts as "active" the same way every time.
Use a consistent invite rule (so results stay comparable)
- Tie invites to the same event definition: for example, "ticket closed" not "ticket updated."
- Cap frequency per customer: starter default-adjust after your baseline and contact patterns: no more than 1 invite per customer per 30 days for transactional surveys; increase the cap if customers have frequent contacts.
- Avoid only sampling edge cases: do not limit invites to escalations, refunds, or VIPs unless that is the exact population you want to manage.
- Keep opt-outs honored: if someone opts out of feedback requests, suppress future invites across channels.
If you will not follow up, do not ask for name, email, phone, or order ID. If you will follow up, ask for contact details only after an explicit opt-in and tell customers who will see their feedback, how long you will keep it, and how you will use it.
Next step: align your survey fields and retention settings with your internal privacy and consent guidance so you are not storing extra PII "just in case."
Frequently Asked Questions
Should I use NPS, CSAT, or CES for customer feedback?
Pick based on the moment you are measuring: use NPS for relationship/loyalty trend tracking, CSAT for satisfaction with a just-finished interaction, and CES for friction in a specific task (checkout, onboarding step, support flow). Keep one primary metric per short survey, then add 1-2 driver questions tied to that touchpoint.
How many questions should a customer feedback survey have?
Internal starter target (adjust after your baseline completion rate): keep transactional surveys to 6-10 questions, and go up to about 12 only when you need light segmentation. Use one primary metric, 1-2 drivers, one open-ended "what should we change?" prompt, and optional segmentation at the end.
When should I send a customer feedback survey?
Internal starter targets (adjust after your baseline): send within 15-60 minutes after support resolution, 1-7 days after confirmed delivery, and right after your onboarding "first value" milestone. If you send reminders, keep it to one reminder max for transactional surveys and use a frequency cap so customers are not over-surveyed.
How many responses do I need for reliable results?
It depends on how you plan to segment results. If you only need an overall trend, start by collecting enough responses to keep the line stable run to run; if you need channel, product, or agent/team cuts, you will need enough responses in each segment to avoid noisy swings.
Should my survey be anonymous or should I collect contact info?
Start with minimal PII by default. Collect contact info only when you will follow up, and always include an opt-in so customers control outreach. Identifiable feedback helps recovery workflows, but anonymity can increase candor, so be explicit about how feedback will be used and who can access it.
Can I compare my NPS/CSAT/CES to industry benchmarks?
Use internal trends first because timing, audience, channel, and question wording change scores. Only benchmark if you can match the other study's setup closely (same touchpoint timing, same scale, similar sampling and channel). If you cannot match methods, treat benchmarks as directional context, not a target.
Related Survey Templates
FREE TO START -- NO CREDIT CARD REQUIRED