Why Customer Satisfaction Score matters
A rising CSAT suggests more respondents rated the experience positively. A decline should be examined alongside survey timing, response volume, comments, and the experience measured.
- Business question
- What percentage of surveyed customers say they were satisfied with this experience?
- Teams that use it
- Customer experience, support, product, operations, and service leadership teams.
- Decisions it supports
- Service coaching, journey improvements, issue prioritisation, and follow-up with dissatisfied customers.
Customer Satisfaction Score formula
Total Positive Responses ÷ Total Responses × 100
Formula components
- Positive responses
- Ratings defined in advance as satisfied, such as 4 or 5 on a five-point scale.
- Total responses
- All valid answers to the satisfaction question, including neutral and negative answers.
- Survey context
- The touchpoint and timing to which respondents are reacting.
How to calculate Customer Satisfaction Score
- Define the survey scale and which answers count as positive before collecting results.
- Remove invalid or duplicate responses using a consistent rule.
- Count positive responses and all valid responses.
- Divide positives by total responses and multiply by 100.
Customer Satisfaction Score example
After support cases close, 500 customers answer a five-point survey. The company defines ratings of 4 or 5 as positive, and 410 responses meet that rule.
- Positive responses = 410.
- Total valid responses = 500.
- CSAT = 410 ÷ 500 × 100 = 82%.
Customer Satisfaction Score is 82%.
Eighty-two percent of people who answered the post-support survey reported a positive level of satisfaction.
How to interpret the result
CSAT is most useful when the question, scale, touchpoint, and timing stay consistent. Segment results by issue type or journey stage to find actionable differences.
Benchmarks vary by survey design, industry, channel, culture, and experience type. A transactional support score should not be compared casually with a relationship survey.
Common mistakes and limitations
- Changing the positive threshold
- Counting only 5s in one period and 4s plus 5s in another breaks the trend.
- Survey response bias
- Customers who answer may have stronger views than those who do not.
- Combining unlike touchpoints
- Checkout, delivery, and support satisfaction answer different questions.
- Reading the percentage without comments
- The score shows how many were positive, not what should be fixed.