● Marketing
CSAT (Customer Satisfaction Score)
CSAT, or customer satisfaction score, measures how satisfied customers are with one specific experience, such as a purchase, a delivery or a support call. Customers rate it on a short scale, usually 1 to 5. The score is the percentage of answers that are 4 or 5.
Full definition
How to calculate CSAT
CSAT is the number of satisfied answers divided by the number of all answers, times 100. This is the standard CSAT formula: CSAT (%) = satisfied responses / total responses x 100. On a 5-point scale, satisfied means a 4 or a 5.
Say 100 customers answer your survey. 80 of them pick a 4 or a 5, and the other 20 pick a 1, a 2 or a 3. Your CSAT is 80 / 100 x 100 = 80%. Only the 4s and 5s sit on top of the fraction, but every answer counts in the total.
CSAT scales and the top-2-box rule
The top-2-box rule counts only the two highest answers as satisfied. On a 1 to 5 scale, that means 4 and 5. A 3 is neutral, so it does not count as satisfied.
Other scales work too. On a 1 to 10 scale, you decide which answers count, for example the top two or the top three. Smileys and thumbs follow the same logic: count the positive answers and divide by all answers. Pick one rule and keep it. If you change the scale or the rule, you can no longer compare the new score with the old ones.
When to send a CSAT survey
Send it right after the interaction you want to measure. A CSAT survey scores one moment, so it works best while the memory is fresh. Good triggers are a delivery, a closed support ticket, an appointment or a meal.
Timing changes what you measure. A survey sent on the day of purchase scores the checkout and nothing else. A survey sent three weeks after a delivery scores a blurred memory. Send one survey per interaction, and avoid asking the same person several times in a row.
What is a good CSAT score
A good CSAT score is one that beats your own earlier results on the same question, with the same scale and the same moment. Published benchmarks mix industries, channels and scales, so there is no single target worth quoting. A restaurant that surveys guests after dinner and a software company that surveys users after a support ticket measure very different things.
Start with a baseline. Collect answers for a few weeks, note the score, then watch the trend. A move from 78% to 84% after you fix late deliveries tells you more than any sector figure.
Read the low answers too. The score tells you how many customers are unhappy. The open comment tells you why. If most 1s, 2s and 3s point to the same step, that step is your first fix. Compare channels with care as well: an in-person survey and an email survey reach different people, so the same business can score differently on each.
CSAT vs NPS vs CES
The three scores answer three different questions.
- CSAT asks how satisfied the customer was with one experience. Use it right after a purchase, a delivery or a support call.
- NPS asks how likely the customer is to recommend you, from 0 to 10. It measures the overall relationship, not one moment.
- CES asks how easy it was to get something done, such as solving a problem. It fits support and returns.
Many teams use CSAT after key moments and NPS once or twice a year. For the bigger picture, read our guide on customer satisfaction, then see how to combine NPS and customer reviews.
Limits of CSAT
CSAT only reflects the customers who answer. Very happy and very unhappy customers often answer more than the ones in the middle, so the score can lean one way.
It also scores one moment. A customer can rate a delivery 5 and still leave next month because of the product. A bare score does not say what to change either, so add one open question. Finally, people read scales differently from one country to another, so compare scores from the same audience.
Concrete example
These four examples are illustrative, with made-up counts.
Restaurant. A restaurant texts guests a one-question survey after the meal. 60 guests answer and 51 pick a 4 or a 5. CSAT is 51 / 60 = 85%. The 9 other answers come with comments about waiting time.
B2B SaaS. A software company sends a survey when a support ticket closes. Out of 40 answers, 30 are a 4 or a 5, so CSAT is 75%. The team reads the 10 low answers and sees that most mention a slow first reply.
Clinic. A clinic asks patients how satisfied they were with their appointment. 120 patients answer and 108 give a 4 or a 5. That is 108 / 120 = 90%. The reception team reads the open comments about check-in.
Online store. A store sends the survey after delivery. 200 customers answer and 150 pick a 4 or a 5, so CSAT is 75%. Most of the 50 low answers mention a late parcel, so the store looks at its carrier first.
Each score only means something next to the same business's earlier results. An 85% for a restaurant and a 75% for a store are not a ranking.
With Review Collect
CSAT needs answers, and answers need a request at the right moment. Review Collect sends post-purchase requests by email, SMS or WhatsApp. On average, 40% of customers respond on SMS and WhatsApp, and merchants collect 30 times more reviews in the first 30 days.
Once reviews are published, the AI replies to all of them in under 60 seconds, positive and negative. To measure CSAT after each order, see NPS and CSAT tracking.
