CSAT (Customer Satisfaction Score), Explained

CSAT, customer satisfaction score, is the most common way support teams put a number on how a customer felt about an interaction. After a conversation closes, the customer is asked a short question such as "How would you rate the support you received?" and answers on a small scale. CSAT is the share of those responses that fall on the positive end.

The classic form asks for a 1-to-5 rating and counts the 4s and 5s as "satisfied"; the score is satisfied responses divided by total responses, expressed as a percentage. Other teams use a thumbs-up/thumbs-down, a 1-to-3 scale, or a set of emoji. All of them are CSAT, and because the scale and the cutoff differ, two CSAT numbers are only comparable when they were collected the same way.

What CSAT does and doesn't tell you

CSAT measures satisfaction with a specific interaction, at the moment it ends. That makes it a fast, local signal: it tells you whether this conversation went well. But it is a narrow one. It is not a measure of loyalty (that is closer to NPS) or of effort (closer to CES), and it says nothing about the customers who never answered. Response rates on post-chat surveys are usually low, and the people who respond skew toward the extremes, so a raw CSAT can flatter or punish a team unfairly when the response base is small.

Reading it honestly

There is no universal "good" CSAT score to target, and quoting an industry-average figure is misleading because scale, wording, timing, and audience all move it. The number worth watching is your own CSAT, measured consistently and read as a trend rather than an absolute. Segment it by channel, by topic, by whether AI or a human resolved the conversation, because an aggregate score hides the interactions that are failing. And always weigh it against response rate: a high score from a handful of replies is noise, not evidence.

CSAT in Evoriqa

Evoriqa collects CSAT inside the same inbox where conversations are resolved, so satisfaction is captured on both AI-handled and human-handled interactions rather than bolted on afterward. It appears on the analytics dashboard next to resolution and deflection, which is where it earns its keep: it is the honesty check on those efficiency metrics, catching the case where automation looks productive but customers left unhappy. Because the agent answers from your knowledge and hands off cleanly when a person is needed, CSAT reflects the quality of the answer and the handoff — not a bot that closed conversations by wearing customers down.