Measuring Customer Satisfaction: NPS, CSAT and CES, and When to Use Each

Why isn’t “Are you satisfied?” enough?

Many organisations measure customer satisfaction, but few use the result to make a decision. Often the question does not match the goal: you ask about overall satisfaction when you really want to know why customers call three times to complete one transaction. Each of the three common metrics asks a different question, and understanding the difference is the first step.

CSAT: rating a specific touchpoint

CSAT asks about satisfaction with a particular service or interaction, usually on a 1–5 scale, and is often reported as the share of customers choosing the top two ratings. Its strength is that it is direct, easy to understand, and tied to a recent experience the customer remembers.

But high satisfaction does not necessarily mean loyalty. A customer may be pleased with a polite employee and still leave frustrated because the transaction took three visits. CSAT is useful for rating channels, branches and staff, and needs an open question about the reason.

CES: was the service easy?

CES asks how easy it was to get the request done. The idea was introduced in a well-known 2010 Harvard Business Review article, based on research with more than 75,000 customers who used contact centres or self-service channels. The authors found that going above and beyond to delight customers makes little difference to loyalty, and that most people simply want a quick, easy resolution. ^[2]

Gartner states that service interactions are about four times more likely to drive disloyalty than loyalty, and lists common sources of effort: repeat contacts about the same issue, switching channels, being transferred, repeating information and complex policies. ^[3]

NPS: the overall relationship

NPS asks one question: “How likely are you to recommend us to a friend or colleague?” on a 0–10 scale. Respondents are grouped as promoters (9–10), passives (7–8) and detractors (0–6). The score is the percentage of promoters minus the percentage of detractors, so it ranges from −100 to +100. ^[1]

A worked example: of 200 respondents, 110 are promoters (55%), 50 are passives and 40 are detractors (20%). NPS = 55 − 20 = 35.

NPS suits measuring the relationship over a longer period, so it is usually asked periodically rather than after every transaction. The Net Promoter System recommends tracking it regularly and breaking it down by segment, business unit, product and region. ^[1] In government services, “recommend” can sound odd when the customer has no alternative; there, CES and CSAT are more telling.

When should you use each metric?

• Service centre or counter: CES or CSAT straight after the transaction, with “What is the main reason for your rating?”
• Portal or app: CES after self-service is completed, because it reveals complicated steps.
• Service contract or long-term client relationship: NPS quarterly or half-yearly, with interviews with detractors.
• New service launch: CES and CSAT in the first weeks, then NPS once it has settled.

Do not combine all three metrics in one survey after every transaction; a long survey lowers the response rate and makes the sample less representative.

Mistakes that make the numbers misleading

• Sample bias: surveying only completed transactions hides the voices of people who gave up halfway.
• Late surveys: asking two weeks later measures memory, not experience.
• Tying rewards to the number alone: this can push staff to ask for high ratings instead of solving problems.
• Unfair comparisons: comparing a branch handling complex transactions with one handling simple ones without adjusting.
• Ignoring open text: the number tells you there is a problem; the comment tells you where.

From measurement to improvement: close the loop

The most important step after measuring is closing the loop: contacting dissatisfied customers quickly, understanding what happened and resolving their case where possible. The team then groups comment themes weekly, fixes the most frequent ones, and tells customers what changed because of their feedback.

A hypothetical example: a government service centre

A service centre measures CSAT after every transaction and the score stays high, while written complaints keep rising. The centre adds a CES question, which shows that property transfer transactions are the hardest. Comments reveal that customers are asked for a document not listed on the website, so they have to come back.

After updating the document list on the website and sending it by text message when an appointment is booked, the centre tracks the effort score for that transaction monthly to confirm the improvement holds. The figures and details are hypothetical, to illustrate the method.

Conclusion

Choose the metric by what you want to improve: CSAT for a touchpoint, CES for ease of service, NPS for the overall relationship. Keep every survey short with a question about the reason, and close the loop with the customer; measurement that leads to no change teaches people not to answer.