CSAT vs NPS vs CES: Which Customer Metric Should You Use?

Author: Inputly

What is the difference between CSAT, NPS and CES?

CSAT, NPS and CES answer three different questions. CSAT asks how satisfied someone was with one specific interaction, such as a support reply or a delivery. NPS asks how likely someone is to recommend you, which stands in for the overall relationship. CES asks how much effort a task took, such as finding a setting or fixing a problem. They are not three versions of the same number. A customer can rate one support conversation highly, rate the product as a whole poorly, and still fight your admin panel every week. Choosing between them depends on which decision you are about to make, not on which metric is best in general.

| Metric | Question it answers | Typical scale | Best moment to send | | --- | --- | --- | --- | | CSAT | How satisfied were you with this interaction? | 1 to 5, or a word scale | Right after support, onboarding or a delivery | | NPS | How likely are you to recommend us? | 0 to 10 | Monthly or quarterly, to the whole base | | CES | How easy was it to get this done? | 1 to 7, easy to hard | After a task, a flow or a self-serve setup |

TL;DR: Use CSAT for moments, NPS for the relationship and CES for friction inside a task. Pick one primary metric per decision and read the other two as context, never as a tiebreak on the same question.

How is each score calculated?

• CSAT: count the positive answers, usually 4 and 5 on a five point scale, divide by all responses and multiply by 100. Keep the threshold fixed, because moving it breaks the trend.
• NPS: subtract the share of detractors, who answered 0 to 6, from the share of promoters, who answered 9 or 10. Passives from 7 to 8 sit in the denominator and never appear in the score.
• CES: average the effort ratings, or report the share of respondents who chose the two easiest points. Either way, keep the direction fixed so lower always means less effort.

NPS can be negative, and that is normal for young products rather than a sign that the survey is broken.

Which metric should you use?

Match the metric to the decision in front of you.

| Your decision | Primary metric | Why | | --- | --- | --- | | Did that support process change help? | CSAT | The moment just happened, so the memory is fresh | | Is the product relationship improving this quarter? | NPS | It tracks willingness to recommend over time | | Where does the setup flow lose people? | CES | Effort points at friction you can remove | | Which theme deserves a roadmap slot? | Feedback volume, then CES | Demand plus friction outweighs one rating |

For a small team, one primary metric plus one open question beats a full dashboard. Add the others only when someone will actually read them every week.

Why do the three scores disagree?

They measure different time horizons, and that is the point. NPS is a slow relationship signal that barely moves after one release. CSAT is a fast reaction to one interaction and swings with a single bad ticket. CES diagnoses one task, so it can fall while CSAT rises if support is charming and the product is exhausting.

The other reason is who answers. A survey sent to everyone reads differently from one sent to people who just finished a workflow, and self-selection shapes the number as much as the experience does, the trap described in survey bias. Read every score next to its response count, and check how many survey responses you need before treating a small sample as a trend.

How do you run all three without exhausting your users?

Do not stack them in one survey. Each metric asks the respondent to switch context, and every extra question costs completion rate, the effect covered in how to reduce survey fatigue. A workable rhythm:

1. One CSAT question at the end of every support conversation and onboarding sequence, sent automatically.
2. One CES question inside the product after someone finishes a setup or import flow, shown once, not on every visit.
3. One NPS question monthly or quarterly to the whole base, with a single open follow-up.

That is three touches per customer per quarter at most, each attached to something they already did. Keep the wording neutral, as any customer feedback survey should, then route the open answers into the same inbox so they can be counted with feedback analytics instead of lost in a report.

What does a good score look like?

Above 80 percent is strong for CSAT, an NPS above 30 is healthy for most SaaS products, and a CES where most respondents pick the easiest two points is the goal. Treat all three as context: a free tool and a premium service do not share a baseline.

The number you can manage is your own trend, segmented by plan, region or workflow. A stable CSAT near 78 percent every week tells you more than one 92 percent from a happy week. When a score drops, split it by segment: a concentrated drop means a bug or a policy change, a broad one points at positioning or pricing. If churn is already moving, read how to reduce customer churn next, because the fix there is rarely a better survey.

Frequently Asked Questions

1. Can one survey measure all three? It can, but the completion rate drops with every added question. Keep one metric per moment and run the relationship question on its own rhythm.
2. Should the scores be anonymous? Offer the choice. Named responses let you follow up on a low score, anonymous ones surface the criticism people only give when nobody is watching, as how to collect anonymous feedback explains.
3. Do you need a tool to run these surveys? No, but you need one place where the answers land. An in-app feedback widget plus a single inbox covers the first year for most teams.
4. Which score belongs in the board deck? One, and the same one every quarter. Switching to whichever number improved is the fastest way to lose trust in all of them.
5. How do you act on a low score? Read the open comments, group them, and close the loop with the people who wrote them, the routine described in how to close the feedback loop.

Pick your primary metric this week and stop debating the other two. Create a free space on Inputly to run CSAT, CES and NPS surveys in one place, tag the answers, and show customers what changed.