VoC vs CSAT vs NPS vs CES: Which Customer Feedback Metric to Track in 2026

The short version: CSAT measures satisfaction with one interaction, NPS measures overall loyalty and likelihood to recommend, CES measures how hard a task was, and VoC is the program that pulls them all into one feedback system. They are not interchangeable. Each answers a different question. The mistake most small teams make is picking one in isolation and either over-using it or under-using the others.

This guide breaks down all four in plain language, then gives you a small-team-friendly recommendation for which combination to use and when. By the end you will know whether to send a CSAT survey after every ticket, whether NPS belongs in your quarterly review, and whether CES is worth the effort.

The one-sentence version of each metric

Metric	What it measures	When to send	Scale
CSAT	Satisfaction with one specific interaction	Right after a ticket, call, or purchase	1–5 or 1–7
NPS	Overall loyalty and willingness to recommend	Quarterly or after a major lifecycle event	0–10
CES	How easy it was to get something done	Right after a workflow (sign-up, support resolution)	1–5 or 1–7
VoC	The whole feedback program — not a metric	All the time, across every channel	n/a

VoC is the operating model. CSAT, NPS, and CES are the three numbers most often used inside a VoC program. Picking between them is really a question of which interaction you are measuring and what decision the number is going to inform.

Customer Satisfaction Score (CSAT)

CSAT asks one question: "How satisfied were you with [the thing that just happened]?" Customers rate on a scale, usually 1–5. The score is the percentage of customers who picked 4 or 5.

When to use CSAT

• After a support ticket closes
• After a phone call
• After a purchase or check-out
• After any single, bounded interaction

CSAT's strength

It is direct, fast, and easy for customers to answer. Response rates are usually higher than NPS because the question feels low-effort. The number maps cleanly to specific touchpoints, which makes it useful for coaching individual agents and spotting product areas with rising friction.

CSAT's weakness

It is a snapshot of one moment, not a trend in loyalty. A customer can give 5 stars after every ticket and still churn at renewal because the underlying product gap was never solved by support. CSAT also suffers from a known ceiling effect — survey responders skew positive, so a 4.6 average can hide a real problem under the average.

CSAT example

"How satisfied were you with the support you received today? 1 (very unsatisfied) to 5 (very satisfied)."

Some teams add an open-text follow-up: "What is the one thing we could have done better?" The verbatim answers are often more useful than the number itself.

Net Promoter Score (NPS)

NPS asks: "How likely are you to recommend us to a friend or colleague?" Customers rate 0–10. Scores split into three buckets:

• 9–10: Promoters — loyal, will refer
• 7–8: Passives — satisfied but not enthusiastic
• 0–6: Detractors — at risk, may speak negatively

The score is the percentage of promoters minus the percentage of detractors. Scores can range from −100 to +100. Anything above zero is technically net positive. 30+ is good. 50+ is excellent.

When to use NPS

• Quarterly across the whole customer base
• After a major lifecycle event (90-day-in, post-renewal)
• As a top-line board metric

NPS's strength

It is the cleanest single number for tracking brand-level loyalty over time. It correlates with retention, lifetime value, and word-of-mouth growth. Investors and execs understand it because it has been the standard CX metric since 2003.

NPS's weakness

It does not tell you why the number changed. A drop from 42 to 28 over a quarter could be product, pricing, a billing issue, a launch that disappointed, a competitor running ads, or all five. NPS needs paired follow-up questions to be useful as a diagnostic tool.

NPS also responds slowly. You will not see the impact of a fix until the next quarterly survey lands. That makes it a poor tool for fast iteration.

NPS example

"On a scale of 0 to 10, how likely are you to recommend [Company] to a friend or colleague?"

Follow-up: "What is the main reason for your score?"

The open-text follow-up is where most of the value lives. The number tells you the trend. The verbatims tell you the cause.

Customer Effort Score (CES)

CES asks: "How easy was it to [complete this task]?" Customers rate on a 1–5 or 1–7 scale, sometimes with the wording "How much effort did you have to put in?" The lower the score, the better.

When to use CES

• After a self-service workflow (sign-up, account setup, ticket submission)
• After a support resolution
• After a checkout
• Anywhere you suspect friction

CES's strength

CES is the metric that maps most directly to product friction. A high effort score on the sign-up flow tells you exactly where to invest engineering time. CES is also more predictive of churn than CSAT, according to a well-cited Harvard Business Review study from 2010 that introduced the metric.

CES's weakness

It is narrower than NPS or CSAT. CES only makes sense when the customer just completed a task. It does not work as a top-line loyalty metric or a brand-health signal. Many teams over-apply it and get diminishing returns.

CES example

"How easy was it to resolve your issue today? 1 (very difficult) to 7 (very easy)."

Or the original 2010 phrasing:

"The company made it easy for me to handle my issue." (Strongly disagree to strongly agree, 1–7)

Voice of the Customer (VoC)

VoC is not a metric. It is the program that captures, analyses, distributes, and acts on customer feedback across every channel. CSAT, NPS, and CES are three of the signals that flow through a VoC program. So are open-text comments, support tickets, reviews, and behaviour-based churn data.

If you want the full breakdown, see our pillar guide: what is Voice of the Customer?.

When VoC sits at the top of all three

The most useful framing for a small team is this: VoC is the system, the three metrics are the dials. You do not run a VoC program by tracking only NPS. You do not run it by tracking only CSAT. You pick one number for each decision the program needs to make:

• Day-to-day quality: CSAT after every ticket
• Friction in a specific workflow: CES after that workflow
• Quarterly brand health: NPS every 90 days
• Patterns and themes: tagged feedback volume (the unsung fourth metric)

Together they make the dashboard. Alone they each leave blind spots.

Side-by-side comparison

Question	CSAT	NPS	CES
What does it measure?	Satisfaction with one moment	Long-term loyalty	Ease of completing a task
When do you send it?	Right after the interaction	Quarterly or post-lifecycle	Right after a workflow
What is a "good" score?	75–85% positive	30+ good, 50+ excellent	Below 3 (on 1–7) excellent
What is its biggest weakness?	Ceiling effect, narrow scope	Slow signal, no cause	Only works post-workflow
Who should own it?	Support / CX	Exec / leadership	Product / ops
Best for fast iteration?	Yes	No	Yes
Best for board reporting?	Sometimes	Yes	Rarely

For deeper coverage of each metric (calculation formulas, benchmarks by industry, when each was introduced), see our pillar on customer satisfaction metrics every team should track.

Which combination should a small team actually use?

Most small support teams should run CSAT continuously + NPS quarterly + CES selectively + open-text everywhere. Concretely:

1. CSAT after every closed ticket

Single 1–5 question with one optional open-text field. The number tracks quality of support. The verbatim flags individual issues. Aim for 75%+ satisfied (4 or 5) within 60 days of launching, then track monthly.

2. NPS once a quarter

A 90-day cycle is long enough to see real movement and short enough that the data stays fresh. Use the standard 0–10 wording and always include the "main reason for your score" open-text follow-up.

3. CES on one or two specific workflows

Pick the workflow you suspect has friction. Sign-up. Self-service password reset. Submitting a ticket. Track CES for two months. If the number moves with product changes, keep the survey. If nothing changes, drop it.

4. Open-text feedback everywhere

The numbers are the trend lines. The open-text is the cause. Always include at least one open-ended question with every survey, and read 20 verbatims per month. This is the cheapest source of insight in any feedback program.

For practical wording examples across all three metrics, see our Voice of the Customer survey: 25 questions that get real answers guide.

Common mistakes when picking a metric

Tracking NPS as a daily metric. NPS does not move daily. Sample-size noise will swamp any signal. Use it quarterly, not daily.

Using CSAT as your only number. CSAT has a ceiling effect — most respondents skew positive. A 4.6 average can hide a churn cliff. Pair it with NPS for the long-term view.

Skipping CES because "we already have CSAT." CES and CSAT measure different things. A workflow can score 5 on CSAT (because the agent was friendly) and 2 on CES (because the customer spent 40 minutes finding the right form). The combination tells you what the average hides.

Treating the score as the goal. The score is the diagnostic. The goal is the decision the score informs. Boards that ask "what is our NPS?" are asking the wrong question. The right question is: "what did we change last quarter, and did the number move with it?"

Sending surveys without acting on results. Customers stop responding to surveys when nothing visibly changes. The fastest way to kill a VoC program is to collect feedback nobody acts on. The fix is the discipline of closing the customer feedback loop — going back to the specific person with what you heard and what you did.

Picking a tool before defining the program. A VoC platform without a clear program generates feedback nobody reads. Define what you are trying to learn first, then pick from our Voice of the Customer tools for small support teams guide.

Frequently asked questions

Is NPS or CSAT better?

Neither is better in isolation. They answer different questions. CSAT tells you whether the last interaction went well. NPS tells you whether the customer is likely to stay and refer. Most mature teams run both: CSAT for day-to-day quality and NPS for quarterly brand-health. Picking one means accepting a blind spot in the other dimension.

What is the difference between CSAT and NPS?

CSAT measures satisfaction with one specific interaction (a ticket, a purchase, a call) on a 1–5 scale and reports the percentage of 4s and 5s. NPS measures overall loyalty across the whole relationship on a 0–10 scale and reports promoters minus detractors. CSAT is a snapshot. NPS is a trend.

What is the difference between CSAT and CES?

Both measure a single interaction, but they ask different things. CSAT asks "were you satisfied?" CES asks "how hard was it?" A customer can be satisfied with the outcome and still rate the effort poorly — for example, after spending an hour navigating a confusing flow before reaching a friendly agent. CES is closer to a friction metric than a satisfaction metric.

Where does VoC fit alongside CSAT, NPS, and CES?

VoC is the program. CSAT, NPS, and CES are three of the inputs to that program. A working VoC program will track all three (plus open-text feedback, support tags, and review data) and roll them into a monthly review that drives at least one decision per cycle.

What is the best survey metric for a small team?

If forced to pick only one, CSAT after every closed support interaction is the most useful single number for a small team. It is fast to set up, has high response rates, and traces directly to specific reps and product areas. Add NPS once you have 90 days of CSAT data and want a brand-level view.

How often should you send each survey?

CSAT after every closed ticket. NPS quarterly. CES after the specific workflow you are diagnosing, then either keep it (if it moves with product changes) or drop it. Avoid sending the same customer more than one survey per week or NPS more than once per 90 days.

What is a good benchmark score for each?

CSAT: 75–85% positive is good, 85%+ is excellent. NPS: 30+ is good, 50+ is excellent, anything above zero is net positive. CES: on a 1–7 scale, anything under 3 is excellent (low effort). Benchmarks vary heavily by industry, so internal trend matters more than absolute level.

Picking your starting point

If your team has no feedback program today, start with CSAT after every closed ticket and read the verbatims weekly for 30 days. Add NPS quarterly once CSAT is steady. Add CES only when you suspect a specific workflow has friction worth measuring.

Each metric on its own leaves blind spots. The combination is what builds a real Voice of the Customer program. For the broader picture of how the three metrics fit into a working VoC system, see our customer satisfaction metrics deep dive. Before you wire up any of them, it is worth writing the one-page customer feedback strategy that says which metric decides what — without it, the numbers stack up but the decisions do not.