CSAT (Customer Satisfaction Score) is an index that measures customer satisfaction after a specific interaction — a call, a chat conversation, an order. Guests were asked "How satisfied are you?" and scored (usually 1-5 stars), the results are converted into a percentage of satisfaction: CSAT = (number of customers scoring 4-5) ÷ (total number of responses) × 100. Unlike NPS (measuring loyalty) and CES (measuring ease), CSAT measures at-the-moment emotions — and is the easiest index for small businesses to start with.
TL;DR
- CSAT = percentage of customers satisfied after an interaction, asked on a scale of 1-5.
- Formula: (customers score 4-5) ÷ (total responses) × 100.
- Different NPS (loyalty) and CES (ease) — measure three different things.
- Increase safety: quick response, correct resolution the first time, consistent attitude, closing the loop after complaints.
- Simple collection: send survey immediately after interaction via Zalo/email.
CSAT calculation formula and how to read the results
The formula for calculating CSAT is the percentage of customers scoring satisfaction (4-5) out of the total number of responses received:
`
CSAT (%) = (Number of customers giving 4-5 stars) ÷ (Total number of responses) × 100
`
For example: 100 customers answered the survey, 82 scored 4-5 → CSAT = 82%.
How to read: There is no absolute "how much is good" threshold, but:
Over 80%: good for most industries.
70-80%: good, with room for improvement.
Below 70%: need to review the process — maybe a system problem, not just staff attitude.
More important than the absolute number is the trend — the decrease in CSAT over the weeks is an early warning signal that action is needed, even though there has not been a major incident.
How to collect CSAT?
The way to collect CSAT is to ask immediately after the interaction ends — the closer to the time, the more accurate the results. Three popular ways:
Survey in the chat channel — after the agent closes the conversation, send a 1-5 star question right in Zalo/Facebook/livechat. Highest response rate because customers are still there.
Email survey — link sent after order delivery or support completion. Suitable when customers are familiar with email.
Post-call score — requires the customer to press the rating key after ending the call.
Sample note: You should only ask after an interaction with a clear result (conversation closed/order/problem resolved), and keep the question short so as not to upset the customer. A response rate of 30-40% is acceptable for a chat channel.
How are CSAT, NPS and CES different?
CSAT, NPS and CES measure three different aspects — should be used in conjunction, not interchangeably.
Index | What to measure | Question | Scale
CSAT | Satisfaction after specific interaction | "Are you satisfied with this support?" | 1-5 stars
NPS | Loyalty/willingness to recommend | "Likely to recommend to others?" | 0-10
CES | Ease of using the product/service | "How easy is it to solve the problem?" | 1-5 or 1-7
For example, customers may be "satisfied" (high CSAT) but not loyal enough to recommend (low NPS). Small businesses should start with CSAT because it is easy to collect and directly contact with daily service quality; When there is enough data, add NPS periodically.
5 ways to increase CSAT for small teams
Improving CSAT is improving two things: response speed and resolution quality — both of which can be done with good processes, without the need for more people.
Faster response — FRT under 15 minutes with chat channel; Slow news is the biggest source of frustration.
First time correct resolution (FCR) — train agents to ask for enough information right from the beginning, avoiding customers having to repeat.
Consistency across channels — customers who text Zalo and then call are still handled the same way (read: What is Omnichannel?).
Close the post-complaint loop — when the issue is resolved, confirm with the customer again before closing — "done, please check".
Tie CSAT to data — compare low scores with each agent/each conversation to train the right people, on the right problem.
Data: Zenify supports automatic scoring (AI QA) for 100% of conversations on the system, helping the team detect low scores immediately instead of waiting for periodic surveys.
Mistakes when using CSAT
Only ask when the customer is comfortable — surveying only after a good interaction distorts the data; must ask randomly.
Only look at the numbers, not the reasons — low scores without comments don't know what to do about it.
Compare CSAT with competitors using different samples — each place asks differently; Comparing with yourself over time is meaningful.
Inaction — measuring without improving is just a "nice number on the dashboard".