NPS vs CSAT vs CES: Which Metric Should You Use, and When?

Three ways to measure the voice of the customer, each answering a different question at a different moment. A practical guide to when to use each, how to calculate it, and the mistakes that ruin the results.

PeekAI · · Voice of the Customer

In many contact centers, the same question comes up in the management meeting: "What's our customer satisfaction number this month?" Then everyone discovers that each department measures something different under a different name. Marketing tracks NPS, the contact center tracks CSAT, and the digital team has heard about CES and wants to add it.

Having several metrics isn't the problem. The problem is that each one answers a different question, and using it in the wrong place gives you a number that is arithmetically correct but useless for making decisions.

Net Promoter Score (NPS): will the customer recommend you?

Fred Reichheld introduced NPS in his well-known 2003 Harvard Business Review article, developed with Bain & Company. It rests on one question: "How likely is it that you would recommend us to a friend or colleague?" on a 0 to 10 scale.

How it's calculated, as Bain explains it:

  • Scores of 9 or 10 are promoters.
  • Scores of 7 or 8 are passives and are left out of the calculation.
  • Scores of 0 to 6 are detractors.

NPS = % promoters − % detractors. With 50% promoters and 20% detractors, your score is +30. The range runs from −100 to +100.

When to use it: to measure the overall relationship with your brand, ideally on a regular cycle (quarterly, for example) rather than after a specific interaction. It's a strategic metric that belongs in the boardroom more than in a team leader's scorecard.

A warning: don't hold contact center agents accountable for post-call NPS. A customer may hate the refund policy or the plan price, and no agent can change that.

Customer Satisfaction (CSAT): was this interaction good?

The oldest and simplest of the three. Right after a specific interaction, you ask: "How satisfied are you with the service you received today?", usually on a 1 to 5 scale. The result is typically the share of respondents who chose 4 or 5.

When to use it: immediately after a call, chat or ticket closure. It's the best fit for evaluating agents, teams and channels because it's tied to a clear moment.

Its weakness: it captures a feeling in the moment. A customer can be pleased with a friendly agent while the problem stays unsolved and they call back tomorrow.

Customer Effort Score (CES): how hard was it to get the problem solved?

CES came from research published in Harvard Business Review in 2010, "Stop Trying to Delight Your Customers." The researchers analyzed more than 75,000 service interactions and reached a conclusion that changed how many companies think: 96% of customers with high-effort interactions became more disloyal, compared with only 9% of those with low-effort experiences.

The message is clear: loyalty doesn't come from delighting customers. It comes from reducing the work they have to do to solve their problem: calling back, being transferred, repeating information, switching channels.

The most common version today is a statement customers rate, "The company made it easy for me to handle my issue," on a 1 to 7 agreement scale. As MeasuringU notes, the wording changed between the first and second versions, so don't compare your results with published figures until you've confirmed they use the same wording.

When to use it: after service journeys that should be easy: updating details, filing a complaint, activating a service, or using self-service.

The comparison in one table

NPSCSATCES
QuestionWould you recommend us?Were you satisfied with this interaction?Was it easy to solve your issue?
LevelBrand relationshipA specific interactionA specific journey or process
TimingPeriodic (e.g. quarterly)Right after the interactionAfter the process is complete
Best forPredicting loyalty and growthEvaluating agents, teams, channelsFinding friction in processes
OwnerSenior leadershipContact center manager and team leadersOperations and service design

Five mistakes that ruin your results

  1. Measuring everything with one metric. NPS after every call mixes agent performance with the customer's view of the whole company.
  2. Collecting the score without an open question. Always add something like "What's the main reason for your score?" The number tells you where the problem is; the comment tells you what it is.
  3. Ignoring non-responders. The angriest customers may never answer. Track response rates and compare them across channels.
  4. Comparing unlike numbers. Your score against another industry, another country or a different wording isn't a fair comparison. Benchmark against yourself over time first.
  5. Measuring without closing the loop. If nobody contacts detractors and nothing changes in operations, customers will stop answering your surveys.

What this means for your team

Don't pick one metric. Use all three, each in its place: NPS for the relationship, CSAT for the interaction, CES for how easy the process is. And before any survey, start with one question: what decision will we make based on this number?

Browse free courses

Sources and references

  1. Harvard Business Review: The One Number You Need to Grow (Reichheld, 2003)
  2. Bain & Company: Measuring Your Net Promoter Score
  3. Harvard Business Review: Stop Trying to Delight Your Customers (Dixon, Freeman, Toman, 2010)
  4. MeasuringU: 10 Things to Know about the Customer Effort Score