CSAT (Customer Satisfaction Score) is the percentage of survey responses classified as positive. The usual formula is (positive responses ÷ all responses) × 100. It measures a specific interaction, not overall loyalty. In Intercom, Amazing and Great count as positive; Okay and Bad are neutral; Terrible is negative.
This page is the definition, formula, and Intercom operating guide. For the other end of the same survey, see CSAT vs DSAT. For targets, see Intercom CSAT benchmarks. For pulling the report, see how to build a CSAT report in Intercom.
Surveyed CSAT only covers customers who rate the conversation. Supportman posts every Intercom rating to Slack as it lands and scores unrated threads with IQS, so the weekly review is not limited to the people who clicked an emoji.
What is CSAT?
A CSAT survey asks a customer how satisfied they were with one conversation, order, or touchpoint. Teams collect a rating plus an optional comment, then roll ratings into a percentage they can trend week over week.
CSAT is useful because it is tied to a specific event. A customer can love the product and still rate support poorly after a refund policy, a slow first response, or a Fin handoff that missed the question. Read the remark before treating the score as an agent grade.
The score only represents people who answered. Unrated conversations are invisible unless you add a second layer such as Intercom's CX Score, a QA sample, or predicted satisfaction on closed threads. See what 100% CSAT coverage means.
How CSAT is calculated
CSAT (%) = (number of positive responses ÷ total responses) × 100
On a classic 1–5 scale, positive usually means 4 and 5. Neutral (3) and negative (1–2) stay in the denominator. They are not “positive,” so they lower CSAT even when they are not counted as DSAT.
| Responses | Count | In CSAT numerator? |
|---|---|---|
| Positive | 82 | Yes |
| Neutral | 11 | No |
| Negative | 7 | No |
| Total | 100 | — |
Worked example: 82 ÷ 100 × 100 = 82% CSAT. DSAT on the same set is 7%. Neutral is 11%. The two percentages do not add to 100%.
Always store the response count and response rate next to the percentage. 82% from 100 ratings is a different signal from 82% from 11 ratings.
How Intercom measures CSAT
Intercom conversation ratings use five emoji choices, not numbered stars. Per Intercom's conversation rating documentation:
| Customer choice | Intercom class | Typical ops use |
|---|---|---|
| Amazing, Great | Positive | Counts toward CSAT; coaching examples |
| Okay, Bad | Neutral | In the CSAT denominator; some teams still review Bad comments |
| Terrible | Negative | DSAT / recovery queue |
Intercom CSAT is positive responses divided by all rating responses. Find the score in Reports → Surveyed CSAT. CX Score is a separate Pro add-on metric that estimates satisfaction beyond respondents; do not mix the two numbers in one target.
Your Slack alert rule can be broader than Intercom's negative class. Many teams route Terrible and Bad together. Label that policy so the dashboard definition and the alert queue are not confused. Setup: send Intercom ratings to Slack.
CSAT vs DSAT, NPS, and CES
| Metric | Question it answers | When to use it |
|---|---|---|
| CSAT | Were respondents satisfied with this interaction? | Weekly ops, teammate coaching, channel mix |
| DSAT | Which responses were negative? | Recovery queue and root-cause grouping |
| NPS | Would they recommend the company? | Product and brand loyalty, not ticket coaching |
| CES | How hard was it to get the job done? | Complex B2B workflows; often overlaps FCR plus reopen analysis |
Start with CSAT and DSAT on the support queue. Add NPS when leadership wants a loyalty number. Keep them on separate review cadences so a relationship survey does not drown conversation-level action.
What is a good CSAT score?
There is no universal CSAT number that fits every queue. Channel, intent mix, survey timing, and response rate all move the result. Use your own rolling baseline first, then compare like-for-like segments.
For Intercom teams, a practical operating band from published customer examples is in the Intercom CSAT benchmark framework: treat 85%+ as strong and 90%+ as excellent only when the mix of chat vs email and Fin vs human is stable. Do not copy a call-center industry average onto an Intercom inbox.
A high CSAT with a collapsing response rate is not a win. A stable CSAT with rising DSAT count is a work queue. Track both.
Run CSAT as an ops metric
- Define positive and negative. For Intercom, start with the vendor classes above, then document any extra ratings you route as alerts.
- Review weekly, not only monthly. Headline CSAT, rating count, DSAT count, and one recovered conversation. Template: weekly customer support report.
- Segment before you average. Chat vs email, Fin vs human, and teammate. Blended CSAT hides the queue that is slipping.
- Close the loop on negatives. Own, classify, recover if a next step exists. Playbook: respond to negative CSAT.
- Cover unrated conversations separately. QA sample, CX Score, or predicted CSAT / IQS. Surveyed CSAT is the respondent slice, not the whole week.
CSAT is a company metric that happens to land in the support inbox. Product bugs, billing policy, and marketing promises show up in remarks. Tag root cause so coaching does not treat every Terrible as an agent miss. For the rest of the scorecard, use customer service metrics.
Frequently asked questions
What does CSAT stand for?
CSAT stands for Customer Satisfaction Score. It is the share of survey responses classified as positive, usually reported as a percentage.
How do you calculate CSAT?
Divide the number of positive responses by the total number of responses, then multiply by 100. Neutral and negative responses stay in the denominator. Do not divide by total conversations unless you have a separate coverage metric.
What is a good CSAT score for support?
Compare against your own trend and channel mix. Intercom does not publish a single platform benchmark. See the Intercom CSAT benchmarks article for operating bands from published customer examples.
How is CSAT different from NPS?
CSAT asks about a specific interaction. NPS asks whether the customer would recommend the company. A customer can give a high CSAT on a refund conversation and still be a detractor on NPS, or the reverse.
Does Intercom use a 1–5 CSAT scale?
No. Intercom uses five emoji labels. Amazing and Great are positive, Okay and Bad are neutral, and Terrible is negative. CSAT is positive divided by all ratings.