Tips
What is customer effort score? How CES works
13 Min Read
Customer effort score measures how easy or difficult customers find a specific interaction. Learn how CES surveys work, how to calculate the score, when to ask for feedback, and how to turn difficult experiences into practical improvements.
Share article:
CES measures perceived effort for a particular task or interaction, such as resolving a support issue or finding an answer.
Choose a clearly labeled scale and document its direction: higher scores do not always mean a better experience.
Calculate the average from valid ratings, show the response count, and inspect the distribution rather than relying on the headline alone.
Ask close to the experience, read customers' explanations, and investigate repeated contacts, handoffs, unclear steps, and missing guidance.
Customer effort score measures how easy or difficult customers find a specific interaction. Learn how CES surveys work, how to calculate the score, when to ask for feedback, and how to turn difficult experiences into practical improvements.
What customer effort score measures
Customer effort score, usually shortened to CES, is a survey measure of how much work a customer feels they had to do during a specific experience. In customer service, that experience might be getting an account unlocked, correcting an invoice, or understanding how to use a feature.
The question concerns the customer's experience of the process. A quick reply does not necessarily make the process easy if the customer must repeat information, move between channels, or work out the next step alone. Equally, a technically complicated request can feel manageable when ownership and instructions are clear.
Qualtrics describes CES as a measure attached to an interaction, including getting a question answered or a request fulfilled. That focus makes it useful for investigating a particular support journey rather than asking customers to summarize their entire relationship with the business.
Imagine two customers who both regain access to their accounts. One follows a clear recovery guide and completes the task. The other tries several articles, contacts support twice, and repeats the same details to a second agent. Both cases can end as resolved, but the effort behind those outcomes is different.
CES, satisfaction, and loyalty answer different questions
CES asks about ease or effort. Customer satisfaction, or CSAT, asks how satisfied someone was with an experience. Net Promoter Score, or NPS, asks about likelihood to recommend. They provide different perspectives and should not be treated as interchangeable scores.
A customer may appreciate a helpful agent while finding the company's approval process exhausting. A high satisfaction rating for the agent would not erase that friction. Similarly, an easy password reset does not establish that the customer is happy with the product overall.
Use CES alongside the broader customer service metrics your team already reviews. The purpose is to understand a part of the experience that speed, closure status, or satisfaction alone can miss.
Choose a customer effort score scale you can explain
There is no useful interpretation of a bare score such as “CES: 5” without knowing the question and its labels. Survey formats vary, including five-point and seven-point scales, direct ease questions, and statements customers rate by agreement.
For this article's examples, use a seven-point ease scale: 1 means very difficult and 7 means very easy. Higher scores indicate an easier experience. A different survey might ask how much effort was required, with 1 meaning very little effort and 7 meaning very high effort. On that survey, lower is better.
Survey format | How to read the result |
|---|---|
Ease question: 1 = very difficult, 7 = very easy | Higher ratings indicate an easier experience. |
Effort question: 1 = very little effort, 7 = very high effort | Lower ratings indicate less effort. |
Agreement with a statement that the task was easy | Greater agreement indicates an easier experience; preserve the exact labels and numerical coding. |
Pick one format for a defined measurement and retain it across reporting periods. Show the endpoint labels beside the score in dashboards and exports. Keep intermediate labels consistent too, especially when translating the survey into another language.
Write a specific, neutral question
A practical question is: “How easy or difficult was it to resolve your account access issue?” It names the task without assuming that the interaction was good. If resolution is uncertain, ask about getting help with the issue instead, or separately establish whether it was resolved.
Avoid combining several experiences in one rating. “How easy was it to find an answer, contact support, and complete setup?” leaves you unable to tell which part the customer evaluated. Choose the decision you need to inform, then define the experience around it.
An agreement format is another option. SurveyMonkey's CES templates describe a seven-point agreement scale and pairing the rating with a follow-up question. Whatever format you choose, check that customers can understand it without an explanation from an agent.
Add one optional open-text question, such as “What made this easy or difficult?” Keep it neutral so customers can describe a helpful part of the experience as well as a problem. Test the survey on mobile and with the assistive technology relevant to your audience; labels should not depend on color alone.
How to calculate customer effort score
A straightforward CES calculation is the arithmetic mean of valid numerical ratings collected with the same survey definition:
Average CES = sum of valid ratings ÷ number of valid ratings
Suppose five customers answer the seven-point ease question with ratings of 6, 5, 7, 4, and 3. Their ratings add up to 25. Divide 25 by five valid responses to get an average of 5.0 out of 7. Under this article's labels, higher means easier. These numbers illustrate the calculation and are not a benchmark.
Count submitted ratings, not survey invitations, in this denominator. If you send 100 invitations and receive 20 valid ratings, calculate the mean from those 20 ratings. Report the invitation and response counts separately so readers can see how much of the eligible group answered.
Do not treat a skipped question as zero. Zero is outside the example scale and would distort the mean. Define how you handle duplicate submissions, test records, and incomplete responses before reporting trends. A genuine negative rating is not invalid simply because it is inconvenient.
Label percentage-based reports separately
Some teams also report the percentage of responses meeting a chosen ease threshold. For example, a team might classify ratings of 6 or 7 as favorable on its seven-point ease scale. If 12 of 20 valid ratings meet that definition, the favorable share is 60%.
That percentage is a different summary from the average. State the threshold and denominator rather than labeling both measures simply “CES.” A 60% favorable share and a mean of 5.0 out of 7 cannot be compared as though they used the same units.
Choose the definition before reviewing whether a change looks successful. Moving the favorable threshold from 6 to 5 changes the measurement; it does not demonstrate that customers encountered less friction.
Keep the distribution visible
An average can hide very different experiences. On the example ease scale, five ratings of 5 have the same mean as ratings of 1, 3, 7, 7, and 7. The second group includes customers who found the experience difficult, even though the headline average is identical.
Show how responses are distributed across the scale, together with the number of ratings. If you designate a difficult-experience band, document it and keep it stable. Review the comments in that band instead of assuming the mean describes every customer.
Do not pool raw ratings from different scale lengths or opposite directions. Reversing a seven-point code with 8 minus the original rating can align numerical direction, but it does not make different question wording or sampling methods equivalent. If you change the survey, record the change and establish a new baseline when comparisons would be misleading.
When to ask customers for a CES rating
Ask close enough to the experience that customers can remember the work involved. Decide what event triggers the survey, who is eligible, and how long after that event the invitation appears. Apply that rule consistently rather than letting agents choose only their smoothest conversations.
For a support journey, the trigger might be completion of an interaction or a defined case-resolution event. For self-service, it might follow an article visit and task attempt. For a product workflow, it might follow a specific action such as updating billing details. The wording should match the event being measured.
Do not confuse a closed ticket with a resolved issue
A system can close a ticket automatically after inactivity. That does not establish that the customer completed the task. If a survey asks about resolution after every administrative closure, some customers will reasonably respond that the premise is wrong.
Check how the trigger behaves for abandoned conversations, merged tickets, reopened cases, and multi-step investigations. An interaction-level survey can ask about getting help so far. A resolution-level survey needs a credible resolution event or a separate outcome question.
Keep those views distinct in reporting. Customers still waiting for a result can reveal important friction even when they are not eligible for a survey specifically about a completed resolution.
Include the people your trigger might miss
A survey displayed only after successful checkout excludes customers who could not finish. A support survey reaches people who contacted the team, but not necessarily those who gave up trying to find the contact option. Recognize those limits when describing the result.
Use other evidence to investigate missing journeys: abandonment data where available, failed searches, repeated attempts, support comments, and customer interviews. An exit or attempt-based survey can help in some contexts, but it measures a different population from a completion-based survey and should be labeled accordingly.
Set a reasonable contact-frequency rule so frequent users are not asked after every small action. Record how sampling and suppression work. If you change the invitation channel, delay, or eligible group, consider whether the change could explain the trend before attributing it to service improvements.
Identify what makes an experience difficult
CES tells you where to investigate. The explanation usually comes from the customer's comment and the underlying journey. Start by reading a sample of difficult-experience responses alongside the relevant conversation or task history, using only the access and customer data needed for the review.
Look for recurring forms of work: explaining the same issue again, trying several channels, waiting without a promised update, repeating steps that already failed, or interpreting instructions that assume missing knowledge. Code the themes consistently so a routing problem does not get grouped vaguely with every other complaint.
Separate the effort source from the final person who handled the request. An agent might provide a careful answer after the customer has already experienced a broken form, several transfers, and an unclear policy. The useful response may require a product or process change rather than individual coaching.
Follow a concrete high-effort journey
Consider a hypothetical customer who needs to change the billing contact. They search for “invoice email,” open an article titled with an unfamiliar internal term, discover halfway through that an administrator must make the change, and contact support. The first reply repeats the article without addressing their permissions.
The work is now specific enough to address. The documentation owner can use the customer's language in the title and explain prerequisites earlier. Support can distinguish missing permissions from an ordinary how-to question. The product team can consider whether the interface explains who can complete the action.
Helpview's guide to content gaps in your help center offers more detail on investigating missing or hard-to-find guidance. The CES comment gives a starting point; journey review establishes what actually happened.
Turn findings into a small improvement backlog
Group repeated patterns, estimate how many customers encounter them, and consider severity as well as volume. A rare account-recovery barrier may deserve attention before a common minor wording issue. Assign a named owner and a review date to each chosen change.
For teams writing customer guidance in Notion, a Notion help center can provide a place to publish task-focused answers. Review prerequisites, decision points, error explanations, and next steps before adding more content. Publishing another article will not help if the existing answer is correct but customers cannot recognize it in search.
Check help-content changes against knowledge base analytics, relevant ticket themes, and customer comments. Article views alone cannot show that a task became easier. A useful result is that customers can find the answer, apply it, and continue without avoidable repetition.
Interpret CES without overclaiming
There is no single target that fits every CES question and support journey. Establish your own baseline with stable wording, labels, timing, and eligibility. Compare similar tasks and customer groups before using a movement in the overall score to judge an improvement.
Report the question, scale direction, calculation, reporting period, valid response count, and response rate where the invitation denominator is known. Include the distribution and the most useful comment themes. If a subgroup contains only a handful of ratings, treat it as a signal for investigation rather than a reliable performance ranking.
Watch for changes in who answers
Respondents are not automatically representative of every customer. People with very positive or negative experiences may be more willing to answer. A new survey channel can change who sees the question; a product release can change the mix of requests reaching support.
For example, an improved billing workflow might receive a lower overall ease score during a week dominated by unusually complex cases. Compare like-for-like issue groups and inspect their counts before concluding that the change made the experience worse.
CES is also a reported perception, not a stopwatch or a complete prediction of retention. A relationship between effort and later behavior does not prove that changing the survey score causes that behavior. If the business question concerns repeat use, review that outcome directly alongside the relevant customer experience.
Check the result after a change
For the billing-contact example, retain the same survey definition while improving the article and support guidance. Review later comparable responses, comments about permissions, repeated contacts, and whether the task was completed. Record other changes that could affect the comparison.
Do not improve the number by making support harder to reach, selecting only resolved cases without disclosing the change, or asking customers to choose a favorable rating. The aim is to remove unnecessary work while preserving access to help and any required security checks.
Conclusion
Customer effort score helps teams understand how easy or difficult a specific experience felt to the customer. Its value depends on a clear question, a consistently labeled scale, a transparent calculation, and a fair survey trigger. Read the score alongside its distribution and customer explanations, then investigate the handoffs, instructions, and repeated work behind difficult journeys. Use those findings to make a focused change and check whether comparable customers experience less effort afterward.
Frequently asked questions
What is customer effort score in customer service?
Customer effort score is a survey measure of how easy or difficult customers found a particular experience, such as getting help or resolving an issue. It captures their perceived effort rather than simply measuring reply speed, ticket closure, or satisfaction with an agent.
How do you calculate customer effort score?
Is a higher or lower customer effort score better?
When should you send a CES survey?
What is a good customer effort score?
Share article:
2 free months of Pro
Turn Notion pages into help center answers.
Keep writing in Notion and publish a real, searchable Notion help center.
Articles
Keep reading
