Ask five people which customer metric matters most and you'll get five different answers, usually because they're each optimizing for something different. NPS, CSAT, and CES aren't competing standards — they measure different things, at different points in the customer relationship. Picking the wrong one for the wrong moment is the most common mistake we see, and "NPS vs CSAT vs CES" is close to a false question: the honest answer for most teams is that you eventually want all three, asked at different times, for different reasons.
This guide covers what each metric actually measures, when to use which one, how CES specifically compares to NPS and CSAT (a common point of confusion), and how to set all three up without reinventing the question wording yourself.
Net Promoter Score (NPS): long-term loyalty
The question: "On a scale of 0–10, how likely are you to recommend us to a friend or colleague?"
Respondents split into three groups: detractors (0–6), passives (7–8), and promoters (9–10). Your score is the percentage of promoters minus the percentage of detractors — so it can range from -100 to +100. A score of +30 or higher is generally considered good; above +50 is excellent — though benchmarks vary heavily by industry, so compare yourself to your own history and close competitors more than to a generic number.
Use it when you want a single, trackable number for overall brand health, measured over time — quarterly or after major product changes, not after every interaction. It's a relationship metric, not a transaction metric: it asks about the customer's overall impression of you, not their experience with any one interaction.
Don't use it when you want to know why someone is unhappy about a specific support ticket or a specific checkout flow — NPS is too coarse for that. Pair it with an open-ended follow-up ("what's the primary reason for your score?") or you'll have a number with no explanation attached, which is the single most common NPS mistake: teams report the score in a dashboard and never learn why it moved.
Customer Satisfaction (CSAT): moment-in-time sentiment
The question: "How satisfied were you with [this purchase / this support interaction / this feature]?" — usually a 1–5 star or numeric scale.
Use it when you want to measure a specific touchpoint right after it happens: right after a support ticket closes, right after checkout, right after onboarding. CSAT is a snapshot, and it decays fast — ask it too long after the interaction and you're measuring memory, not experience. Send it within minutes to hours of the interaction, not days later.
Don't use it when you're trying to capture how someone feels about your company as a whole. A great CSAT score on one support ticket says nothing about whether that same customer would recommend you to anyone — someone can be perfectly satisfied with a single resolved ticket and still be quietly shopping for alternatives because of unrelated friction elsewhere in the product.
Customer Effort Score (CES): how hard was this
The question: "How easy was it to [resolve your issue / complete this task]?" — typically a 1–7 agreement scale ("strongly disagree" to "strongly agree" with "this was easy").
CES exists because satisfaction and effort turned out to be different things: a customer can be satisfied with the outcome of a support interaction while still finding the process frustrating (three transfers, two repeated explanations). CES specifically isolates that friction, which CSAT and NPS both miss — a customer can rate a resolved ticket 5/5 for satisfaction and still churn six months later because every interaction with you costs more effort than it should.
Use it when you're trying to reduce churn driven by process friction — support, returns, account changes, anything with steps a customer has to get through. CES is the strongest predictor of repurchase intent of the three metrics in the original research that introduced it, specifically because low-effort experiences build loyalty even when nothing about the outcome itself was exceptional.
Customer effort vs. customer satisfaction: what's the actual difference?
CSAT asks "were you happy with this," CES asks "how hard was this to get through." They're correlated but not the same thing, and the gap between them is exactly where the useful signal lives. A customer effort vs customer satisfaction comparison in practice usually shows the two moving together most of the time — but the interactions worth paying attention to are the ones where they diverge: high satisfaction with high effort (a customer glad they got the outcome but exhausted by the process — churn risk you won't see in CSAT alone) or low satisfaction with low effort (the process was smooth, but the outcome itself was the problem — a product or policy issue, not a process one). If you only track CSAT, that divergence is invisible.
Customer effort vs. NPS: two different distances
Customer effort vs NPS is a different kind of comparison, because they're not really measuring adjacent things — NPS is a relationship-level judgment made over months or years of accumulated experience; CES is a reaction to one specific interaction. A single high-effort support call is unlikely to move your NPS by itself, but a pattern of high-effort interactions absolutely will, over time. Think of CES as an early-warning input into NPS rather than a competing metric: if your CES scores are trending down across support interactions, expect NPS to follow eventually, just with a lag.
A worked example: a SaaS company across the customer lifecycle
Say you run a mid-size SaaS product. Here's roughly when each metric earns its place:
- Onboarding (first 14 days): CES, right after the setup flow — "how easy was it to get your account configured?" This is a high-effort moment by nature (new users, new interface), and effort here predicts whether they come back at all.
- First support ticket: CSAT immediately after the ticket closes — did this specific interaction go well?
- Quarterly, to your full active user base: NPS — the overall relationship check-in, independent of any one interaction, that tells you whether the accumulated experience is trending toward advocacy or churn.
- Cancellation flow: CES again — "how easy was it to cancel," which sounds like an odd thing to optimize, but a high-effort cancellation flow doesn't retain unhappy customers, it just converts them into unhappy customers who also complain publicly about how hard you made it to leave.
Notice that no single metric covers the whole lifecycle, and none of them substitute for the others — that's the actual answer to "NPS vs CSAT vs CES": they're not alternatives, they're different instruments pointed at different distances.
Comparison at a glance
| NPS | CSAT | CES | |
|---|---|---|---|
| Question | "How likely are you to recommend us?" | "How satisfied were you with X?" | "How easy was it to do X?" |
| Scale | 0–10 | 1–5 (star or numeric) | 1–7 agreement scale |
| When to ask | Quarterly, or after major milestones | Immediately after a specific interaction | Immediately after a specific task or process |
| What it tells you | Overall relationship health, long-term loyalty | Sentiment about one touchpoint | How much friction a process created |
| Time horizon | Long-term (relationship) | Short-term (transaction) | Short-term (transaction) |
| Best predictor of | Advocacy, brand health | Immediate satisfaction | Repurchase intent, churn from friction |
| Common mistake | No open-ended follow-up to explain the score | Asking too long after the interaction | Treating a high-effort outcome as acceptable because satisfaction was still high |
Picking one (or more)
A rough rule of thumb:
| If you're measuring... | Use |
|---|---|
| Overall brand loyalty, tracked over time | NPS |
| A specific interaction, right after it happens | CSAT |
| How much effort a process took | CES |
Most teams that run all three use NPS quarterly, CSAT after every support ticket, and CES after specific high-friction flows (returns, cancellations, account recovery) — not the same survey, not the same cadence. Running all three isn't overkill; it's the normal setup for a team that wants both the long-term number (NPS) and the operational signal to actually act on (CSAT, CES).
FAQ
Is CES better than CSAT? Neither is "better" — they answer different questions. CSAT tells you whether someone was happy with an outcome; CES tells you how much friction it took to get there. Use CES specifically when you suspect process friction is costing you customers even though outcomes look fine on paper.
Can you use NPS and CSAT together? Yes, and most mature feedback programs do — NPS as the periodic relationship check-in, CSAT after individual interactions. They're complementary, not redundant: NPS tells you the trend, CSAT tells you which specific interactions are dragging it down.
What's the difference between CES and NPS? CES measures effort on one specific task, asked immediately after it happens. NPS measures overall loyalty across the whole relationship, asked periodically. A single bad CES score won't move NPS by itself, but a pattern of them will, over time.
Do I need all three metrics, or can I just pick one? If you can only run one, NPS gives you the broadest signal for the least survey fatigue — but it also gives you the least actionable detail. Most teams start with NPS, then add CSAT at their highest-volume touchpoint (usually support) once they need to know why the NPS number is moving.
What's a good CSAT score? It depends heavily on the interaction type and industry, but CSAT is usually reported as the percentage of respondents who chose the top one or two options on the scale (e.g. 4 or 5 out of 5). There's no single universal benchmark — track your own trend over time rather than chasing an external number.
Setting these up in Surveyee
All three question types — the 0–10 NPS scale, star-rating CSAT, and Likert-style agreement scales for CES — are built in, and we have ready-to-use templates so you're not building the question wording from scratch:
- Customer Satisfaction template — CSAT plus an NPS follow-up
- NPS Survey template — the standard NPS question with a reason-for-score follow-up
Both are free to use, no account required until you're ready to publish and start collecting real responses. If you're deciding between a 0–10 NPS scale and a shorter CSAT scale for a question you're building from scratch, see our guide on choosing the right rating scale for the granularity tradeoffs.