CSAT score is a post-interaction survey metric measuring how satisfied a customer was with a specific touchpoint. Calculated as (number of satisfied responses ÷ total responses) × 100. Industry average for B2B contact centers is 75–85%.
Definition
Customer Satisfaction Score, usually shortened to CSAT, measures how a customer felt about one specific interaction. That interaction might be a support call, a collections conversation, an insurance enrollment call, or a case follow-up. The point of the metric is not to describe the whole brand relationship. It asks: did this touchpoint meet the customer’s expectation?
That narrow scope is why contact center leaders use CSAT for staffing, coaching, and process reviews. If a queue has weak CSAT, you can usually trace the problem to script quality, long hold times, confusing policies, or a mismatch between what agents promise and what operations can actually deliver.
Formula
Most teams ask customers to rate satisfaction on a 1–5 scale. Ratings of 4 and 5 count as satisfied. Ratings of 1, 2, and 3 do not. Once you have the response count, the math is simple: divide satisfied responses by total responses, then multiply by 100.
Example: if 160 customers answer the survey and 132 choose 4 or 5, CSAT is 82.5%.
How to survey
Question wording
The cleanest CSAT question is direct: How satisfied were you with this interaction? Avoid mixing satisfaction with effort, loyalty, or product quality in the same question. If you ask several things at once, your score becomes hard to interpret. Keep the lead question narrow, then add one optional open-text field if you want detail for QA.
Timing
Send the survey right after the touchpoint or immediately after the case is marked closed. Waiting too long lowers response rate and adds recall bias. In collections and enrollment programs, timing matters even more because customer emotion changes quickly after a payment decision or eligibility answer.
Scale choice
A 1–5 scale is the most common because it is easy for customers to understand and easy for teams to benchmark. Some teams use emoji or text labels, but the underlying idea should stay stable. If you change the wording or the scale every quarter, trend lines stop being useful.
Industry benchmarks
Context matters. A collections team handling delinquent accounts will not score like a B2B SaaS support team handling product questions. What matters is whether you are high or low for your operating model and whether your score moves with changes in quality, wait time, and resolution clarity.
| Performance band | Collections | Insurance enrollment | B2B SaaS support |
|---|---|---|---|
| World-class | >90% | >92% | >90% |
| Good | 80–90% | 82–92% | 80–90% |
| Average | 70–80% | 72–82% | 70–80% |
| Needs improvement | <70% | <72% | <70% |
What drives CSAT in collections and enrollment calls
In collections, customers rarely expect a pleasant call. They do expect a fair one. CSAT usually rises when agents explain options clearly, confirm balances with confidence, avoid repeating questions the customer already answered, and handle hardship conversations with patience. It usually falls when the customer hears vague payment language, gets transferred, or feels trapped by a script that does not match their situation.
In insurance enrollment, the biggest CSAT drivers are different. Customers care about clarity, pace, and confidence in the explanation of plan choices, dates, and next steps. Long hold times or uncertain answers create anxiety because the call often affects care access or cost. That means a team can hit script adherence targets and still hurt CSAT if agents sound unsure or make the enrollment journey feel longer than it needs to be.
CSAT vs. NPS vs. CES
CSAT asks whether a customer was satisfied with one interaction. NPS asks whether the customer would recommend the company at a broader level. CES, or Customer Effort Score, asks how hard the interaction felt. These metrics are related but not interchangeable.
- Use CSAT for queue management, QA reviews, and agent coaching.
- Use NPS for brand loyalty and account-level health.
- Use CES when friction, policy steps, or channel design are the main problem.
A team can have decent CSAT and poor NPS if the single interaction went well but the broader product relationship is weak. A team can also have good CSAT and bad CES if customers feel they had to work too hard to get the answer, even though the agent was polite.
Common mistakes
The most common CSAT mistake is treating the score like a final verdict instead of a sampled signal. Low survey volume, narrow survey timing, and channel bias can distort the number. Another mistake is rewarding agents on raw CSAT without adjusting for case type. Collections hardship calls, cancellation requests, and denied claims will always be harder than password resets or simple billing questions.
Teams also get into trouble when they chase score inflation. If agents push customers for positive ratings, the number may rise for a while but the metric stops helping operations. The goal is not a prettier dashboard. The goal is a tighter link between what customers experience and what managers change.
How AI call scoring predicts CSAT before surveys come in
Most contact centers hear only from a fraction of customers after the call. That leaves long periods where managers are waiting for survey data to tell them what already happened. AI call scoring closes that gap by reading conversation patterns that tend to show up before a low CSAT rating: repeated holds, unclear next steps, unresolved objections, talk-over, or lack of empathy in difficult moments.
The useful part is not prediction for its own sake. It is the ability to find risk early. If a queue’s calls show longer silence gaps, more compliance misses, or more confused customer replies this week, leaders can coach before next week’s surveys land. That is especially useful in low-response environments like collections, where survey lag makes quality issues harder to spot.
FAQ
What is a good CSAT score?
For many B2B contact centers, 80% to 90% is good and more than 90% is top-tier. Below 70% usually means the issue is bigger than one coachable behavior and should trigger a process review.
How often should you survey customers?
Survey after key interactions rather than by calendar. Post-call and post-case-close surveys give the cleanest reading because the experience is recent.
What's the difference between CSAT and NPS?
CSAT measures satisfaction with one interaction. NPS measures likelihood to recommend the company overall. CSAT is better for agent and queue decisions.
Why does my CSAT score fluctuate so much?
Small samples, channel mix, a few very low ratings, and survey timing all create movement. Look at response count and case type before reacting to a short swing.
How do collections teams improve CSAT without reducing recovery rates?
By making the interaction clearer and more respectful, not softer. Better option framing, fewer transfers, and cleaner next-step language can lift CSAT while keeping recovery discipline intact.
See How Your CSAT Compares
Get a clear baseline on satisfaction, handling patterns, and the call behaviors that usually sit behind weak survey results.
Email: amanda@altorlab.xyz
Related metrics: Deflection Rate, Escalation Path, Agent Occupancy Rate, Average Handle Time, First Call Resolution