How to Calculate Customer Effort Score (CES): Formula, Survey and Benchmarks

Customer Effort Score is the average of responses to a single question: "How easy was it to handle your request?" rated on a 1 to 7 agreement scale from strongly disagree to strongly agree. Add up every score and divide by the number of responses. A CES above 5 out of 7 is generally good, above 6 is excellent. The score matters because effort predicts loyalty better than satisfaction does: the harder a customer has to work, the more likely they are to leave.

Customer Effort Score, or CES, came out of research showing that reducing customer effort does more to build loyalty than delighting people does. Most customers do not want to be wowed. They want their problem solved without a fight. CES measures how much of a fight it was. This guide covers the exact formula, the survey question and scale to use, when to send it, what counts as a good score, and, most importantly, how to turn the number into fixes rather than a line on a dashboard.

What is the customer effort score formula?

The formula is a straight average. Take every response on the 1 to 7 scale, sum them, and divide by the total number of responses:

CES = (sum of all effort ratings) ÷ (number of responses)

If ten customers rate their effort 6, 7, 5, 7, 4, 6, 7, 3, 6, 5, the sum is 56 and the CES is 5.6. That is the version most teams use today, because it keeps the full 1 to 7 detail. You will also see an older percentage-based method that reports the share of respondents who "agree" or "strongly agree" (a 5, 6 or 7), stated as a percentage. Both are valid; just pick one and stay consistent, because a 5.6 average and a "72% found it easy" are not comparable to each other.

What survey question should you ask?

The standard modern CES question is a statement the customer agrees or disagrees with: "[Company] made it easy for me to handle my issue." Responses run on a 7-point agreement scale, from 1 (strongly disagree) to 7 (strongly agree). Phrasing it as an agreement statement rather than a "how easy was it?" question is deliberate, because it maps cleanly to a consistent scale and avoids the vague wording that muddies effort questions.

Keep the wording tight and neutral, and always pair it with one open follow-up: "What made it easy or difficult?" The number tells you the effort was high; the open answer tells you it was the three password resets or the transfer between four agents. If you want a deeper primer on wording scales and avoiding leading phrasing, we cover it in how to write good survey questions.

When should you send a CES survey?

CES is transactional by design, so send it immediately after a specific interaction while the effort is fresh: right after a support ticket is resolved, a self-service article is used, an onboarding step is completed, or a purchase goes through. The closer the survey sits to the moment of effort, the more accurate the memory and the higher the response rate. A CES survey sent a week later measures how someone feels about the memory, not the experience.

This is what separates CES from a relationship metric like NPS. You would not send CES quarterly to your whole base; you send it continuously, triggered by the events where effort actually happens. That cadence also means the score moves quickly when you change something, which makes CES a good operational metric for support and product teams working on friction.

What is a good customer effort score?

On the 1 to 7 scale, an average above 5 is generally good, above 6 is excellent, and anything below 4 signals real friction worth investigating now. On the percentage method, roughly 70% or more choosing "easy" is a solid result. Those thresholds are rough guides, not universal truths, because effort norms differ by industry and by the type of interaction: resetting a password should feel effortless, while migrating an account may reasonably take more work.

As with every experience metric, the most useful comparison is against your own trend, not a global number. A CES that slips from 6.1 to 5.4 over a quarter is telling you friction is creeping in somewhere, even though 5.4 still reads as "good." Watch the direction, segment the score by interaction type, and treat a drop as a prompt to go read the open answers behind it.

CES vs NPS vs CSAT: how does effort score fit?

The three metrics answer different questions and work best together. CSAT measures how satisfied someone is with a specific experience. NPS measures overall relationship loyalty and willingness to recommend. CES measures how hard the customer had to work. Effort is the most predictive of the three for repeat behavior in service interactions, because a single high-effort experience does disproportionate damage to loyalty.

Most mature programs run all three at different moments. If you are deciding which to start with, our breakdown of CSAT vs NPS vs CES walks through when to reach for each.

How do you improve your customer effort score?

You improve CES by removing the specific sources of effort your customers keep naming, not by chasing the average. The score is a symptom; the open-text answers are the diagnosis. Read the "what made it difficult?" responses, cluster them into themes, and you will almost always find a short list of repeat offenders: repeating information across agents, unclear error messages, a self-service article that does not match the current product, a handoff that drops context.

Two levers move the number fastest. First, cut the number of steps and handoffs in your highest-volume interactions, so customers resolve in one touch rather than three. Second, make sure the people handling those interactions can actually resolve them, which often comes down to how well you train and onboard your support team so answers are fast and correct the first time. Fix the top themes, watch CES climb over the following weeks, and you will usually see it show up in retention too.

Turning the score into fixes

The gap between measuring CES and improving it is the same gap that stalls most feedback programs: the score gets tracked, the open answers pile up unread, and nobody can say why effort went up. Closing it means treating the verbatim responses as the real data, theming them at scale, and tying each theme to the customers and interactions behind it, so a drop arrives with a named, rankable cause instead of a shrug.

That is what a customer effort score tool does when it connects to the rest of your signals: it reads the open answers, groups them into ranked themes, and joins each to the support tickets and product behavior around it, so you see not just that effort rose but exactly which flow is costing you and how much churn sits behind it. For the broader picture of how effort fits alongside your other feedback, see how to build a voice of customer program. A CES number on its own is trivia. A CES number with the reason attached is a to-do list.