What Is CSAT (Customer Satisfaction Score)?

CSAT is the most direct feedback a contact center gets. The customer has just finished a conversation, and you ask them what they thought of it. Because the question is tied to a single interaction, the answer can be attributed to an agent, a queue, a campaign or a time of day, which is what makes CSAT useful for coaching in a way that broader relationship metrics are not.

The standard form is a five-point scale running from very dissatisfied to very satisfied, and the convention is to count the top two points as satisfied. Some operations use a three-point or seven-point scale, and some count only the top point. None of these is wrong, but all of them produce different numbers from the same customers, so the scale and the counting rule have to travel with the score whenever it is quoted.

The Formula and Its Denominator

CSAT = (satisfied responses / total responses) x 100

The arithmetic is trivial. The interesting part is the denominator, which is responses rather than interactions. If nine hundred calls produced twelve survey responses, a CSAT of 83 percent describes ten people who felt strongly enough to answer. Publishing the response count and the contact count beside the percentage is not a nicety; without it the score cannot be interpreted at all, and quarter-on-quarter comparisons become meaningless when the response rate moves.

Making the Score Worth Acting On

• Ask immediately. The interaction is clearest in the customer's mind in the minutes after it ends.
• Keep it short. One score question and one open comment will out-perform a six-question form on response rate and on honesty.
• Ask in the customer's language. A survey in the wrong language is a survey that quietly selects for bilingual customers.
• Lock the wording. If the question changes, this quarter's score and last quarter's score are answers to different questions.
• Follow up the low scores. A detractor who receives a call back is worth more than a detractor who becomes a data point on a chart.

How DialerBee Handles CSAT

DialerBee's surveys and CX module builds the survey with a live preview beside it, on six mobile-first page templates that carry your tenant branding. CSAT on a one-to-five scale is one of five question types, alongside yes/no, NPS from zero to ten, consent yes/no and open text. When you publish, that version is locked and a change creates a new version, so the answers you collected keep the wording they were collected under.

Delivery covers the channels customers actually use. The survey can run on the call itself as a voice IVR survey answered by keypress or by speech, the Layla survey runtime can ask the questions in conversation on the same call flow, or it can go out as a WhatsApp or SMS message with deliveries logged per channel. Every survey also has a web page behind a short link, and an agent can press Send to survey at the end of a call so it goes out while the conversation is still fresh. The analytics give an overview, a breakdown, a per-question view, a trend and drivers, each stating what the number is out of, plus an export. Low scores and detractors become recovery cases in a queue, assigned to a person and closed with an outcome.

Frequently Asked Questions

How is a CSAT score calculated?

Take the number of responses you count as satisfied, usually the top two points on a five-point scale, divide by the total number of responses, and multiply by 100. A five-point scale where 180 of 250 respondents answered four or five gives a CSAT of 72 percent. The denominator is responses, not contacts, which is why the response rate must be reported beside the score.

What is the difference between CSAT and NPS?

CSAT asks about one specific interaction and is answered immediately after it, so it measures the experience you just delivered. NPS asks about the relationship as a whole and is usually asked periodically. A team can hold a strong CSAT on individual calls while the overall relationship is deteriorating, which is why the two are complementary rather than interchangeable.

When should a CSAT survey be sent?

As close to the interaction as possible, while the conversation is still fresh. Immediately after the call is the usual choice, either on the call itself or as a message sent the moment it ends. Delay costs you both accuracy and response rate, because people forget the detail and lose the motivation to answer.

Why does the response rate matter as much as the score?

Because a score with no denominator is not a score. Eighty percent satisfaction from twelve responses out of nine hundred contacts describes twelve people, and the people who respond are rarely a neutral sample. Always publish the number of responses and the number of contacts alongside the percentage.

Can CSAT be collected during the call itself?

Yes. A voice survey can run at the end of the call, with the customer answering by keypress or by speech, which captures the people who would never open a message afterwards. The trade is length: an on-call survey has to be very short, so it typically carries one score question and perhaps one open comment.

Related terms: NPS, first call resolution, speech analytics, service level and SLA and WhatsApp Business API.