NPS vs CSAT vs CES: which customer metric to use

NPS, CSAT and CES are the three scores most customer feedback programs are built on, and they are often treated as interchangeable. They are not. Each asks a different question, suits a different moment and supports a different decision. Picking the right one, and asking it the same way every time, matters far more than which one is in fashion.

The three at a glance

• Net Promoter Score (NPS) asks how likely someone is to recommend you, on a scale from 0 to 10. People who answer 9 or 10 are promoters, 0 to 6 are detractors, and the score is the percentage of promoters minus the percentage of detractors, anywhere from −100 to +100.
• Customer satisfaction (CSAT) asks how satisfied someone was with a specific experience, most often on a five-point scale. It is commonly reported as the share of people who chose one of the top two answers.
• Customer effort score (CES) asks how easy it was to get something done, such as resolving an issue. A common version asks for agreement, on a seven-point scale, with a statement like “The company made it easy to handle my issue.”

The key difference is scope. NPS is about the relationship as a whole. CSAT and CES are about a moment: a purchase, a support conversation, the first week with a product.

How to ask each one

The wording and the scale are part of each metric, so start from the usual versions and change as little as you can:

• NPS: “How likely are you to recommend [company] to a friend or colleague?”, from 0 to 10, with the ends labeled “Not at all likely” and “Extremely likely”.
• CSAT: “Overall, how satisfied were you with [the experience]?”, on five points from “Very dissatisfied” to “Very satisfied”.
• CES: agreement with “[Company] made it easy to handle my issue”, on seven points from “Strongly disagree” to “Strongly agree”.

The arithmetic is simple enough to check by hand. Say 200 customers answer the NPS question: 90 give a 9 or 10, 70 give a 7 or 8, and 40 give 0 to 6. That is 45% promoters and 20% detractors, so the NPS is 25. The 70 passives count in the total but in neither group. For CSAT, if 120 of 200 people choose one of the top two answers, the CSAT is 60%.

NPS: the relationship

NPS works best as a periodic check on how customers feel about you overall, asked on a regular schedule rather than after every interaction. Its strength is its simplicity: one question, one number that is easy to track, and a split between promoters and detractors that tells you whose comments to read first.

Its weaknesses come from the same simplicity. The score ignores the middle of the scale, so a few people moving from 8 to 9 can shift it noticeably, and on a small sample much of the movement from one quarter to the next is chance. Scores are also hard to compare across companies, industries and ways of asking, so treat published benchmarks with caution. The glossary entry on Net Promoter Score covers what the number can and cannot carry, and the NPS calculator puts a confidence interval around yours.

Keep the schedule independent of what just happened to each customer. If you only ask after support tickets, you only hear from customers who recently had a problem, and the score tracks your ticket queue more than the relationship.

CSAT: a specific moment

CSAT fits right after something happens: an order arrives, a ticket closes, an onboarding call ends. Because it is tied to one moment, it points at something you can fix: this process, this team, this step. Ask while the moment is still fresh; a month later, you are measuring a memory of it.

Its weakness is that there is no single standard version. Five-point and seven-point scales, different labels and different definitions of “satisfied” all produce numbers that look alike and do not compare. Choose one wording and one scale and keep them. Report the full spread of answers alongside the top-two share, because a slide from “Very satisfied” to “Satisfied” will not show in the headline number.

CES: friction

CES asks about effort, which makes it the most operational of the three. It is useful after tasks where ease is the whole point: getting help, changing a plan, finishing a setup. A low score tells you where a process is making customers work too hard.

It says little about value. A customer can find a process easy and still be unhappy with what they got at the end of it. Like CSAT, CES exists in several wordings and on several scales, so only compare it with itself.

Which one to use

Start from the decision you need to make, then pick the score that informs it:

• “Are customers, overall, likely to stay and speak well of us?” Use NPS, asked periodically.
• “Did this interaction go well?” Use CSAT, asked straight after it.
• “Where do our processes make customers work too hard?” Use CES, asked after the task.

Most touchpoints need one of these, not all three. Asking three versions of “how did we do?” in a row lengthens the survey and leaves you with three numbers to reconcile instead of one you understand well.

Whichever you choose, ask why

A score tells you where you stand; the reason tells you what to do next. Follow every score with an open question such as “What is the main reason for your score?”, and read the answers from detractors and dissatisfied customers first. The guide to open-ended questions covers how to turn those answers into themes you can count.

Keep it comparable

All three scores are only useful as trends, and a trend only means something if the measurement stays the same. Keep the wording, the scale, the point in the customer journey and the kind of customer you ask constant. When one of them has to change, note it next to the trend instead of comparing across the change.

Sample size matters too. A score from thirty responses moves a lot by chance, so report how many people answered, and do not act on a small wobble. The margin of error calculator shows how wide the uncertainty is around a percentage such as a top-two share.

Setting it up in Plumeform

In the builder's question picker, “Or insert a validated scale” adds any of the three with standard wording, as ordinary questions you can edit: NPS together with its follow-up reason question, CSAT as a single five-point rating, and CES as a seven-point agreement item. In the results, an NPS question shows the score itself as well as the spread of answers, and rating questions show their average and distribution.

The NPS and customer satisfaction template pairs the recommend question and its reason with a rating of the most recent experience and an ease question. It also shows two questions only to some respondents and repeats the score back in a later question; those two touches need Pro to publish. Remove them, and the rest of the template publishes on Free.