CSAT vs NPS: What Each Score Tells You and When to Use It

Quick Summary

CSAT and NPS measure different things. CSAT (customer satisfaction score) asks how happy a customer was with one interaction, so it shows whether that touchpoint works. NPS (Net Promoter Score) asks how likely they are to recommend you, so it tracks overall loyalty. Most large consumer brands need both:

• Use CSAT right after a key moment, like support, delivery or checkout.
• Use NPS on a regular schedule to spot loyalty risks.
• Use CES (customer effort score) to find friction in service and self-service.

Scores don’t explain themselves. The reasons live in customer comments, so read both together.

Why CSAT and NPS Often Tell Different Stories

Your support CSAT is 87% and steady. Your quarterly NPS just fell nine points. Leadership wants to know which number is right, and what it means for retention.

Both are right. They measure different things, at different moments, from different groups of customers. That's why, at scale, the CSAT vs NPS question rarely has a one-metric answer.

The pressure to get this right is growing. In the American Customer Satisfaction Index's Q1 2026 release, US customer satisfaction slipped to 76.7 out of 100, while complaints rose 16% in a single quarter to a record high. When satisfaction stalls, CX teams need to know which score to trust for which decision, and what is actually moving it.

This guide shows when each score is the right one, where CES fits, and how to use the comments behind both to explain any change. 

Why Listen to Us?

Since 2015, Chattermill has helped high-volume consumer brands make sense of their NPS, CSAT and CES feedback. Uber's work with us started in 2018 as an NPS project in Latin America. It now covers all five of Uber's mega-regions, with more than 400 people across CX, Ops, Product and Support using insights from NPS, sentiment surveys, app reviews and social media.

Footasylum unifies around 2,500 NPS responses a week with its CSAT responses and Trustpilot reviews in Chattermill. It holds an NPS above 80, and grew revenue 7% year on year in 2023. This guide reflects how teams like these read the two scores together.

CSAT vs NPS: The Short Answer

CSAT measures satisfaction with a specific experience. NPS measures loyalty to your brand. Use CSAT when you need to know whether a touchpoint, such as a support contact or a delivery, is working. Use NPS when you need to track the health of the whole customer relationship.

In one line each:

• CSAT is the share of customers who rate an interaction 4 or 5 on a 1 to 5 scale. Our CSAT guide covers the details.
• NPS is the percentage of promoters (9 or 10) minus the percentage of detractors (0 to 6) on a 0 to 10 scale. See our NPS guide for more.

CSAT is short-term and specific. NPS is long-term and broad. That's why they can move in opposite directions, and why a mature CX program runs both.

CSAT vs NPS Side by Side

Here's how the two scores compare.

Treat benchmarks as context, not targets. Retently's figures come from its own customers' surveys, and ACSI reports a modeled 0 to 100 index rather than a percentage, so neither lines up exactly with your internal scores. Region matters too: Retently notes that European customers rate more conservatively than US customers and give fewer 9s and 10s. Your own trend is the benchmark that matters most.

When to Use CSAT vs NPS

When CSAT is the Right Metric

Use CSAT when the question is "did this moment work?" It suits:

• Support teams tracking each channel, queue or agent
• Operations teams checking delivery, returns or store visits
• Product teams testing a new checkout flow or feature release

Because it's sent close to the event, CSAT gives fast, specific feedback that's easy to hand to the team that owns the touchpoint. Its limit is scope. A customer can be satisfied with every interaction and still leave because a competitor is cheaper. Once CSAT comments run into the thousands each week, CSAT analytics tools help find the drivers behind the score.

When NPS is the Right Metric

Use NPS when the question is "how healthy is the relationship?" It suits:

• Leadership reporting and company-level CX goals
• Tracking loyalty across segments, regions or plans over time
• Spotting churn risk before it shows up in revenue

Relationship NPS goes to a sample of your whole customer base on a set schedule. Transactional NPS asks the recommend question after a specific journey, so some brands use it in place of CSAT at key moments. The limit of NPS is that it rarely tells you what to fix. A drop could come from price, product, delivery or a competitor's launch. Teams handling large NPS volumes often use NPS analysis software to read the comments behind every move.

When You Need Both

For a large DTC or B2C brand, the question is which decisions each score should drive. A common split:

• CSAT and CES as operational metrics, owned by the teams that run each touchpoint
• NPS as the strategic metric, owned by CX leadership and reported to the executive team
• Open-text analysis across both, so any change in either score can be traced to a cause

That split also helps when leadership asks how CX affects revenue. NPS shows the relationship trend, while CSAT and the comments show where it's being won or lost.

NPS vs CSAT vs CES: Where Customer Effort Fits

Customer effort score (CES) asks how easy it was to get something done. It's usually phrased as a statement such as "The company made it easy to handle my issue," rated on a 1 to 7 agreement scale, and sent straight after a support contact, self-service journey or return.

CES earns its place because effort is a strong signal of who will leave. In Harvard Business Review's "Stop Trying to Delight Your Customers", researchers at CEB (now part of Gartner) studied more than 75,000 customers and found that effort predicted loyalty better than satisfaction measures or NPS. Gartner's own research finds that service interactions are nearly four times as likely to make customers less loyal as more loyal.

Think of CES as a specialist tool, not a third headline score:

• Use CES when the goal is cutting friction and repeat contacts in service journeys.
• Use CSAT when you need a broader read on how a touchpoint felt.
• Use NPS for the relationship as a whole.

There's no widely used CES benchmark, since companies word and scale the question differently, so track your own trend. Our customer effort score guide covers the common formats. 

How to Read CSAT and NPS Together

Scores tell you that something changed. The open-text comments tell you why. Reading the two together comes down to four habits.

1. Always Ask Why

Pair every CSAT and NPS question with one open-text follow-up, such as "What's the main reason for your score?" Without it, you have a number and no route to a fix.

2. Tag Comments by Theme, with Sentiment for Each

At high volume, nobody can read every comment, and sampling misses quieter issues. Tag each comment against one shared set of themes and score sentiment for each theme separately. A comment that loves the food but hates the delivery should count against delivery, not sit in a "mixed" bucket. Chattermill's Lyra AI uses aspect-based sentiment analysis to pick out several themes in a single comment and the sentiment behind each one.

3. Read the Gap Between The Scores

When CSAT and NPS disagree, the gap is itself a clue:

• High CSAT, falling NPS: Touchpoints work, but something outside them is hurting loyalty, often price, product range or a competitor.
• Falling CSAT, steady NPS: A specific touchpoint is breaking, but loyalty hasn't caught up yet. Fix it before it does.
• Both falling: A broad problem customers feel everywhere, such as an outage or a policy change.
• Both rising: Find which themes improved so you can repeat what worked.

4. Rank Themes by Impact, Not Volume

The most-mentioned theme isn't always the one dragging your score down. Impact analysis estimates how much each theme pushes NPS or CSAT up or down, so you fix what moves the number. Stefan Platteau, Associate Director of Global Product Strategy and Analytics at HelloFresh, says it helps the team see "which themes and categories affect our NPS score," which makes the business case easier to build.

E.ON Next works the same way, ranking themes by a mix of volume and NPS impact in monthly insight reviews. Over the same program, NPS for its billing services rose from +9 to +22.

Worked Example: NPS Falls While CSAT Holds Steady

Here's how this plays out, using the scenario from the start of this guide. The brand and numbers are illustrative.

A meal-kit subscription brand sends relationship NPS each quarter to about 12,000 customers, plus CSAT after every support contact and delivery. Between Q2 and Q3, the scores moved like this:

The touchpoint scores are flat, so the drop isn't coming from support or delivery.

Step 1: Find where the drop sits: Splitting NPS by segment shows the fall is concentrated among subscribers of 12 months or more, whose NPS fell from +44 to +27. Newer customers barely moved.

Step 2: Read the detractor comments by theme: Two themes jumped:

Comments can mention more than one theme, so the columns don't add up to 100%. Typical Q3 comments read like "Price went up again and I'm still getting the same few recipes."

Step 3: Size each theme's impact: Impact analysis attributes about 5 of the 9 lost points to price and value, and 3 to recipe variety. Delivery, the biggest complaint theme in Q2, actually improved, so last quarter's priority list would have pointed the team at the wrong problem.

Step 4: Trace the cause and act: A July price rise had landed alongside a smaller rotating menu. No CSAT survey covers either, so the touchpoint scores never moved. The fix sits with pricing and product: restore menu rotation for long-term subscribers and explain what the new price includes. Next quarter, CX re-measures both themes and the segment's NPS to confirm the fix worked.

CSAT alone would have said everything was fine. NPS alone would have said something was wrong. The comments showed what, for whom and why.

The same pattern shows up in real programs. Outletcity, Europe's largest outlet, spotted negative sentiment about shipping costs in its feedback and introduced free delivery for gold members. That was one of several feedback-led changes behind a 13% rise in NPS, while negative sentiment fell by 33% between 2020 and 2023.

Find Out What's Really Moving Your Scores

CSAT shows whether single moments work. NPS shows whether customers are staying loyal. Neither tells you why, and when you have thousands of comments, reading them all by hand isn’t possible.

Chattermill brings survey answers, support tickets, reviews and social posts together in one place. It tags every comment by topic and feeling, then shows which topics are moving your NPS and CSAT. So when leaders ask what’s behind the numbers, you can answer with proof.

Book a demo to see what’s driving your scores.

FAQs

Is CSAT or NPS Better?

CSAT is better for judging specific touchpoints and giving teams fast, fixable feedback. NPS is better for tracking loyalty and the overall customer relationship. Most large consumer brands use both, with open-text analysis to explain changes in each.

What is the difference between NPS and CSAT?

CSAT measures how satisfied a customer was with one interaction. NPS measures how likely they are to recommend your brand, which reflects loyalty to the whole relationship.

When should you use CSAT vs NPS?

Use CSAT straight after a specific moment, such as a support contact, delivery or checkout. Use NPS on a regular schedule, such as quarterly, to track the relationship. Use transactional NPS when you want the recommend question tied to one journey.

What is the difference between NPS, CSAT and CES?

NPS measures loyalty, CSAT measures satisfaction with an interaction, and CES measures how easy it was to get something done. CES is most useful in service and self-service journeys, where effort is a strong signal of who may leave.

Why is my NPS dropping when CSAT is high?

Usually, the cause sits outside your CSAT surveys. Price rises or a rival’s offer can hurt loyalty without worsening any interaction. To find it, split NPS by segment and sort detractor comments by theme. 

What is a good CSAT and NPS score?

It depends on your industry, region and survey method. Retently's 2026 benchmarks put most industries' CSAT between 65% and 80%, and ecommerce NPS at 61. Treat these as context, and benchmark against your own trend first.