CSAT, NPS, CES or FCR — Which Metric to Use When

Why one metric is not enough

A customer can be happy with a purchase and still never recommend you. They can be loyal, yet every support call costs them half an hour. They can tick "satisfied" on your survey and still switch providers after one painful product return. The four core customer-experience metrics exist because they measure different things — and the most common failure is not bad measurement, but correctly measuring the wrong thing at the wrong place.

CSAT — satisfaction with a specific experience

Customer Satisfaction Score answers "how did this particular interaction go". The classic wording is "How satisfied were you with [the product / service / your experience today]?" on a 5-point scale. The score is the percentage of "satisfied" responses (the top two boxes) out of all responses.

Reference points: 70–80% is good, above 80% excellent, above 90% world-class. The American ACSI index, which tracks 40+ industries, hovers in the mid-to-high 70s. CSAT is strongest immediately after the event — after a purchase, a delivery, a closed ticket, at the venue exit. Its weakness: it measures momentary sentiment, not future behavior. A satisfied customer can still leave.

NPS — loyalty and recommendation

Net Promoter Score asks a different question on a different time horizon: not "how did today go" but "would you recommend us". The global average is around +32 (SurveyMonkey's 2025 benchmark across 150,000+ organizations); above +50 is excellent, above +70 world-class and rare.

NPS is a relationship metric, not a transactional one. Ask quarterly or after key milestones — onboarding completion, contract renewal — not after every interaction. And never without the open follow-up "what is the main reason for your score?" — the number without the reason is a dashboard without a steering wheel.

CES — how much effort it takes to do business with you

Customer Effort Score is the most underrated of the four, and the CEB research (now part of Gartner) is unambiguous: 96% of customers who go through a high-effort service interaction become more disloyal, versus just 9% after a low-effort one. Effort predicts churn better than delight does.

The wording: "[The company] made it easy for me to handle my issue" on a 7-point agreement scale, or more simply — "How easy was it to resolve your issue today?" Reference: a mean above 5 out of 7 is good, above 5.5 strong. CES belongs wherever the customer had to expend effort: after a support call, after self-service, during a product return, at checkout, in an online order flow. It is meaningless as a general "how are we doing" metric — ask only where effort actually occurred.

FCR — did we solve it the first time

First Contact Resolution is simultaneously an experience metric and a cost metric: every repeat call irritates the customer and costs money. The question is binary: "Was your issue resolved during this call?" — yes or no, right after the contact.

SQM Group's benchmarks: the call-center average is 68–70%, 70–79% is good, and 80%+ is world-class — achieved by only about 5% of centers. The practitioner's rule of thumb: every point of FCR improvement brings roughly a point of CSAT improvement and about one percent lower operating cost. If you run a call center, helpdesk, or service desk and don't measure FCR, you are leaving the cheapest improvement on the table.

A word on NSS and the Sean Ellis test

NSS (Net Satisfaction Score) is the net version of CSAT: percent satisfied minus percent dissatisfied. Useful when you don't want neutral answers to blur the picture.

The Sean Ellis test is for product teams: "How would you feel if you could no longer use the product?" If 40% or more answer "very disappointed", you have product-market fit — a threshold derived from benchmarking about a hundred startups, made famous by Superhuman, which grew its score from 22% to 58% using it as the north-star metric.

The decision table

• Relationship-level loyalty / board reporting → NPS (quarterly)
• Quality of a specific transaction or visit → CSAT (immediately after the event)
• Friction in a process — support, returns, ordering → CES (after the process)
• Call center / helpdesk effectiveness → FCR (after the contact) + CES
• Early-stage product → Sean Ellis test (quarterly)

Four common mistakes

• One metric for everything. NPS after every support call measures the wrong thing at the wrong moment — that slot belongs to CSAT, CES, and FCR.
• Mixing scales. A score on a 1–100 scale next to a benchmark on a 0–10 scale in the same table misleads every reader. Compare like with like.
• The number without the "why". Every metric needs one open-ended follow-up. Without it you know that there is a problem, but not which one.
• Over-surveying. A customer asked after every interaction stops responding — or worse, responds on autopilot. One survey per customer every 2–4 weeks is the sensible ceiling.

How kazva.bg covers all four

The kazva.bg platform measures feedback at every touchpoint — a QR code at a physical location, a link after a call, a kiosk at the exit — and computes the scores in real time, segmented by location, period, and channel. The same infrastructure carries the quarterly NPS question, CSAT at the till, CES after support, and FCR after the call — without four separate tools and without manual processing.

Co-authored with Claude Fable 5 (Anthropic).