What Is CSAT? Customer Satisfaction Score Explained

CSAT (Customer Satisfaction Score) is a feedback metric that measures how satisfied someone is with a specific product, service, or interaction, usually by asking "How satisfied were you with [X]?" on a 1-5 scale. Where NPS measures long-term loyalty, CSAT measures how a particular moment went — a support ticket, a delivery, an onboarding call, a training session.

The CSAT question

The standard format asks: "How satisfied were you with your experience?" with answers on a five-point scale, typically labeled from 1 (very dissatisfied) to 5 (very satisfied). Some teams use variants — smiley faces, stars, or a 1-7 scale — but the 1-5 satisfaction scale is the convention that makes results comparable across teams and time.

Because the question names a specific experience, wording matters. "How satisfied were you with today's support conversation?" produces cleaner data than a vague "How satisfied are you with us?", which drifts toward measuring the overall relationship — a job better left to NPS.

How CSAT is calculated

CSAT is expressed as a percentage:

CSAT = (number of 4 and 5 responses ÷ total responses) × 100

Only the top two ratings count as "satisfied." A 3 is treated as neutral, not as a partial win. If 80 out of 100 respondents answer 4 or 5, your CSAT is 80%.

Some teams instead report the average rating (e.g. 4.2 out of 5). That is a legitimate view of the same data, but the percentage-satisfied formula is the standard CSAT definition and is what published benchmarks refer to.

What counts as a good CSAT score?

Commonly cited good ranges fall around 75-85%, but this varies by industry — support interactions, retail, healthcare, and software all benchmark differently, and how and when you ask shifts the numbers too. As with NPS, your own trend line is more useful than an external target: a team that moves from 68% to 78% has a real story regardless of what the industry average is.

Examples of CSAT questions

The pattern is always "How satisfied were you with [the specific thing]?" — the variable is how precisely you name the thing:

• Support: "How satisfied were you with the resolution of your support request?"
• Events and workshops: "How satisfied were you with today's session?"
• Onboarding: "How satisfied were you with the setup process?"
• Purchases and delivery: "How satisfied were you with your order experience?"
• Internal services: "How satisfied were you with the response from the IT helpdesk?"

Each of these works as a one-question survey with an optional "Anything we could improve?" text field behind it.

When to use CSAT

CSAT shines at transactional moments — points where a distinct experience just ended and the memory is fresh:

• Right after a support ticket is resolved.
• At the end of an onboarding step or training session.
• After a purchase, delivery, or appointment.
• At the close of an event or workshop, where a QR code on the final slide catches people before they leave — the guide on QR code feedback for events and booths covers that setup.

Timing is most of the game. A CSAT question asked three days after the interaction measures memory and mood; asked within minutes, it measures the experience. Keeping the survey to one or two questions helps too — a satisfaction score plus one optional comment is usually enough, and Conejo Survey's 30+ survey types include rating scales, emoji reactions, and short-answer follow-ups built for exactly this kind of quick pulse.

CSAT's limitations

CSAT is deliberately narrow, and that cuts both ways:

• It measures a moment, not a relationship. Customers can rate every support ticket 5 and still churn because the product no longer fits their needs.
• It is sensitive to who answers. Very happy and very unhappy people respond more readily, so low response rates skew results.
• A neutral 3 is invisible in the score. A pile of 3s and a pile of 1s can produce the same CSAT, but they describe very different situations — look at the distribution, not just the percentage.

The fix for most of these is pairing the score with an open comment and reading responses as they arrive on a live dashboard, rather than treating the percentage as the whole story.

Improving a low CSAT score

When a touchpoint's CSAT sags, resist the urge to redesign everything at once. Read the 1-2 responses first — dissatisfied customers are usually specific about what went wrong — and look for the moment they cluster around: wait time, a confusing step, a policy, a handoff. Fix the most-mentioned issue, keep the survey identical, and watch whether the next few weeks of responses move. Because CSAT is transactional, it responds to changes much faster than NPS does, which makes it a good metric for verifying that a fix actually landed.

CSAT alongside other metrics

Most feedback programs use CSAT for transactional touchpoints and NPS for periodic relationship checks, with CES (Customer Effort Score) covering how hard things are to get done. They answer different questions and complement rather than replace each other — the full comparison is in NPS vs CSAT vs CES.

Where to go next

• NPS vs CSAT vs CES — pick the right metric for each moment.
• How to write survey questions that get honest answers — the wording rules that keep CSAT data clean.
• What is NPS? — the relationship-level counterpart to CSAT.