CSAT: what the Customer Satisfaction Score is and how to calculate it

CSAT (Customer Satisfaction Score) measures how satisfied a customer was with a specific interaction, product or moment. The classic question is "how satisfied were you?", answered on a scale. It is a point-in-time, transactional metric: it captures the feeling in the heat of the moment, right after a concrete experience, rather than a generic opinion about the brand.

Because it is tied to an event, CSAT answers direct questions: did support resolve the issue? was onboarding smooth? did the new feature land well? Each touchpoint can have its own CSAT, which turns the metric into a magnifying glass over the customer journey, not just an overall scoreboard.

The formula is a simple ratio: the number of satisfied responses divided by the total number of responses, expressed as a percentage.

• CSAT = (satisfied responses / total responses) x 100.
• "Satisfied" is usually the sum of the two highest ratings on the scale, the so-called top-2 box, for example 4 and 5 on a 1-to-5 scale.
• The result reads as "X% of customers were satisfied".

Example: 80 positive responses out of 100 responses give a CSAT of 80%. Defining upfront what counts as "satisfied" is decisive, because changing the cutoff changes the number. The criterion must stay fixed so comparisons across periods make sense.

There is no single scale. The most common runs from 1 to 5, from "very dissatisfied" to "very satisfied", but you also see scales of 1 to 3, 1 to 7, 1 to 10, stars, faces or just a thumbs up and thumbs down.

• 1 to 5: the most used, balancing granularity and simplicity.
• Binary (yes or no, thumbs): fast, great for the end of a chat.
• 1 to 10 or stars: more nuance, but they ask more thought from the respondent.

What matters is standardizing the scale and the "satisfied" cutoff across the whole company. CSAT from different touchpoints is only comparable when the ruler is the same.

All three are experience metrics, but they look at different things. CSAT measures satisfaction with a specific moment, the now. NPS measures loyalty and the willingness to recommend over the long term. CES measures the effort the customer spent to get something done.

• CSAT: point-in-time and transactional, "how satisfied were you?".
• NPS: relational, "how likely are you to recommend us?".
• CES: focused on friction, "how easy was it to solve your problem?".

They are not competitors, they complement each other. A high CSAT right after support living alongside a low NPS, for example, suggests the company resolves isolated cases well but has not yet earned the customer loyalty.

CSAT pays off most when it is triggered right after the experience you want to measure, while the memory is fresh: when a support ticket is closed, when onboarding wraps up, after a purchase or the first use of a new feature.

• Ask at the right moment: minutes or hours after the event, not weeks later.
• Ask a single question, with the scale visible, so you do not tire the respondent.
• Leave an optional open field: the "why" behind the rating is where the action is.

Watch out for sample bias: those who respond tend to be the ones with a very good or very bad experience. Low response volumes make CSAT unstable, so treat a handful of responses with caution before drawing conclusions.

There is no universal magic number. What counts as a good CSAT varies a lot by touchpoint, by industry and by the scale used: the CSAT of a support interaction tends to be higher than that of a billing one, for example. That is why the most useful comparison is with your own history, not with a generic market average.

It is worth remembering the gap between what a company thinks it delivers and what the customer perceives. Bain observed that around 80% of companies believe they deliver a superior experience, while only 8% of customers agree. Tracking CSAT over time, cross-referencing it with Churn and with the Customer health score, turns an isolated rating into a signal of risk or of a healthy relationship.