How to Create a Customer Satisfaction Survey

If you're figuring out how to create a customer satisfaction survey that actually tells you something, the short version is: anchor it to one specific moment, keep the question narrow, and don't let it drift into a general relationship check. A customer satisfaction survey is the fastest way to find out whether a specific experience, a purchase, a support ticket, an onboarding call, landed well. It isn't a loyalty metric and it isn't a product roadmap tool. It's a quick check on one moment, and it works best when you treat it that way. This guide covers the question structure, where to place it, and how to read the score once responses arrive.

By the end you'll have a working CSAT survey, you'll know exactly when to trigger it, and you'll have a plan for what to do with the score once it's calculated.

What CSAT actually measures

Customer Satisfaction Score (CSAT) asks one direct question: "How satisfied were you with [this experience]?" Answers sit on a 1-5 scale, from very dissatisfied to very satisfied. The score is the percentage of respondents who answer 4 or 5 out of everyone who answered.

That's a narrower measurement than it sounds, and the narrowness is the point. CSAT doesn't ask whether someone will recommend you, and it doesn't ask how much effort something took. Those belong to NPS and CES, which get covered in detail in NPS vs CSAT vs CES. CSAT asks only whether this specific thing was good. That specificity is what makes it actionable: you can trace a bad score directly back to the interaction that caused it, which you often can't do with a relationship-level metric.

Step 1: Pick the exact moment you're measuring

CSAT only works when it's anchored to something specific. Before you build the question, decide precisely what "this experience" refers to. It might be the moment right after a support ticket is resolved, right after checkout or delivery, right after an onboarding session, or right after someone uses a new feature for the first time. Any of these works. What doesn't work is leaving it vague.

If you send a CSAT survey "sometime this week" with no trigger tied to a real event, respondents answer based on whatever they remember most recently, and that might have nothing to do with what you're actually trying to measure. Vague timing produces a vague score, every time.

Step 2: Write the question

Keep the wording tied to the moment, not the company in general. "How satisfied were you with the support you received today?" works. "How satisfied were you with your checkout experience?" works. What doesn't work is the generic "How satisfied are you with [Company]?", that drifts toward a relationship metric and loses the precision that makes CSAT useful in the first place.

As with NPS, it's worth adding one optional open-text follow-up: "What's the main reason for your rating?" A 2-star rating with no context tells you something went wrong. A 2-star rating with "the agent didn't resolve my issue and I had to call back" tells you exactly where to fix the process.

Step 3: Build it in SurveyKar

SurveyKar has a dedicated CSAT question type in the rating category, next to star rating, NPS and CES. It renders the standard 1-5 satisfaction scale and calculates the CSAT percentage automatically, so you're not exporting raw numbers into a spreadsheet just to get the score.

Here's how to set it up:

1. Create a new survey and add a question.
2. Select the CSAT question type from the picker.
3. Write the question around the specific moment you identified in Step 1, not a general satisfaction prompt.
4. Add an optional open-text question underneath for the reason behind the score.
5. If the survey follows a support interaction, add conditional logic that routes anyone answering 1 or 2 to a short follow-up asking if they'd like a callback. That turns a low score into a recovery opportunity instead of a dead end.

If you're starting on the free plan, you're in luck here: the CSAT question type is available on SurveyKar's free tier, which makes it an easy first metric to test before you commit to a paid plan. NPS, by contrast, needs a paid plan to use the dedicated question type, which is worth knowing if you're planning to run both.

Step 4: Decide how and when it reaches people

Timing matters as much as wording for CSAT, because the whole point is capturing a reaction while it's still fresh. Send it immediately after the trigger event, within minutes or hours rather than days, a support CSAT sent a week after ticket closure measures memory, not experience.

Where you send it depends on the moment. Embedding the survey in the product or support tool works well right after a ticket closes or a session ends, since it catches the person while they're still in that context. A QR code at a physical point of service, a reception desk or checkout counter, works for in-person moments. And a short web link sent by email or WhatsApp suits post-purchase or post-delivery triggers, especially where WhatsApp is already how you talk to the customer. Full detail on the available channels is on the surveys and feedback page.

Step 5: Set a realistic response expectation and read the score

CSAT surveys are short and situational, so response rates tend to run higher than relationship surveys, often 20-40% for a well-timed, one-question CSAT sent right after an interaction. Still, don't trust a score built on a handful of responses from a busy week. Look at the trend across at least a few dozen responses before you draw conclusions, and run your numbers through the CSAT calculator to see the percentage and check it against a reasonable sample size.

When you read the results, look at the distribution, not just the headline percentage. A CSAT of 75% built mostly from 4s is a very different situation than one built from a bimodal mix of 5s and 1s with few 2s or 3s in between; the second pattern usually means two very different customer experiences are being averaged into one number. Trace low scores back to the trigger event, since that specificity is CSAT's whole advantage, if support CSAT drops in a given week, check staffing or ticket volume before assuming it's a broader problem. And track each touchpoint separately rather than combining them. Onboarding CSAT and support CSAT measure different parts of the business, and merging them into one number hides exactly the detail you're trying to capture.

A worked example

A mid-size e-commerce business sends a CSAT survey immediately after every delivered order, triggered by a WhatsApp message: "How satisfied were you with your delivery?" on a 1-5 scale, with an optional reason field.

In one month, 1,200 orders go out and 340 customers respond, a 28% response rate that's reasonable for a short, well-timed, single-question survey. Of those, 255 rate it 4 or 5. CSAT works out to 255/340 × 100 = 75%.

Reading the open-text answers from the 45 respondents who rated 1 or 2, a clear pattern shows up: most complaints mention late delivery specifically, not product quality. That's an actionable finding a single percentage never would have surfaced on its own, and it points straight at logistics rather than the product team.

Try it

Not sure how many responses you need before the score means anything? Read how many survey responses you actually need.

SurveyKar has a ready-made customer satisfaction survey template with the 1-5 scale and reason field already built in, so you can launch right after your next support ticket closes or order ships. Once responses start coming in, run the numbers through the CSAT calculator to get your score and see how it holds up.

Start measuring customer satisfaction on SurveyKar. Build your CSAT survey from the template, trigger it off the moment that matters, and get a score you can actually act on.