Customer effort score (CES)

Customer effort score (CES) is a single-question survey metric that captures how much effort a customer had to put in to complete an interaction with a business — typically resolving a support issue, completing a purchase, or finding information. It is calculated as the average rating customers give when asked, on a 7-point or 5-point scale, how easy or difficult the interaction was. Lower effort correlates strongly with higher retention and lower churn, which is why CES has become one of the three core post-interaction metrics alongside CSAT and NPS.

Rule of thumb: the modern 7-point CES survey averages ~5.5 across industries. Scores above 6.0 indicate a low-effort experience; scores below 4.5 indicate friction that is likely driving churn. A 1-point improvement in CES typically corresponds to a 12-15 percentage-point increase in repurchase intent.

How customer effort score is calculated

CES is collected through a one-question survey delivered immediately after a defined interaction — most commonly a closed support ticket, a completed checkout, or a self-service session. The customer rates a single statement on a Likert scale. The standard CEB / Gartner version of the question is: "The company made it easy for me to handle my issue."

The customer responds on one of three scales:

• 7-point scale (most common, recommended): 1 = strongly disagree, 7 = strongly agree
• 5-point scale: 1 = very difficult, 5 = very easy
• 3-point scale (legacy): low, medium, high effort

The CES formula: CES equals the sum of all individual responses divided by the total number of responses. For example, if 100 customers respond to a 7-point survey and the responses sum to 540, the CES is 5.4.

Some teams report CES as the percentage of respondents who selected the top two box values (6 or 7 on a 7-point scale). This is sometimes called "% low effort" and is useful for trend reporting, but the raw average remains the standard.

The CES scale, interpreted

On a 7-point scale, scores roughly map to the following experience tiers:

• 6.0 - 7.0: Very low effort. Customers feel the company removed obstacles for them. Strong predictor of repurchase and referral.
• 5.0 - 5.9: Average. Customers got the result they needed but noticed some friction — a hand-off, a repeated question, a slow response.
• 4.0 - 4.9: Elevated effort. Customers had to push to get resolution. This range is correlated with negative word-of-mouth and increased call volume from the same customer.
• Below 4.0: High effort. Strong churn signal. Often indicates broken self-service, repeated escalations, or unclear policies.

CES vs CSAT vs NPS

CES, CSAT, and NPS measure different stages of the customer relationship and are most useful when used together rather than as substitutes for each other.

• CES measures effort during a specific interaction. It is the best predictor of repurchase and loyalty at the transaction level.
• CSAT (customer satisfaction) measures emotional satisfaction with an interaction. It correlates with sentiment but is a weaker predictor of future behavior than CES.
• NPS (net promoter score) measures relationship-level advocacy across the entire brand experience, not a single interaction.

The CEB research underlying CES found that reducing customer effort was a stronger driver of loyalty than "delighting" customers. In other words, removing friction matters more to retention than exceeding expectations.

When to send a CES survey

The timing of the survey shapes the data you get. The two most effective placements are:

• Immediately after a closed support ticket or chat session — captures effort while the interaction is fresh. This is the most common deployment.
• After a self-service interaction — such as a knowledge base article view or completed AI agent conversation, to evaluate whether knowledge base content actually resolved the issue.

Avoid sending CES surveys days after the interaction or bundling them with longer relationship surveys — both significantly reduce response rates and introduce recall bias.

How to act on a low CES

A low CES is a diagnostic signal, not an outcome. The most valuable analysis is segmenting CES by interaction type, channel, and resolution path to find the specific friction points.

• Channel-switching is the single largest driver of high effort. Customers who start in self-service and have to escalate to a human report 2-3x more effort than customers who resolve in a single channel.
• Repeated explanations across hand-offs are the second largest driver. Multi-turn conversation systems that preserve context across agents and channels directly reduce effort.
• Time-to-resolution matters less than perceived effort. A 30-minute interaction with one clear hand-off scores better than a 5-minute interaction with three handoffs.

CES in AI-driven customer service

AI agents change the CES profile in two meaningful ways. First, instant 24/7 availability removes the wait-time component of effort, which is typically the highest-weighted contributor. Second, AI agents that maintain context across turns and channels eliminate the repeated-explanation penalty. In production deployments, fully-resolved AI conversations score 0.5-1.0 points higher in CES than equivalent human-handled tickets, primarily because they remove queueing and hand-off friction.

The risk is the opposite case: AI conversations that fail to resolve and require escalation produce some of the lowest CES scores in any channel, because the customer absorbed the AI effort and then had to start over with a human. Containment rate, deflection rate, and CES should always be monitored together to confirm that contained interactions are actually low-effort, not just contained.

Common pitfalls when measuring CES

• Mixing scales across time — switching from a 5-point to 7-point scale invalidates trend comparisons. Pick one and hold it for at least a year.
• Surveying only resolved tickets — unresolved tickets are precisely where effort is highest and where the data is most diagnostic. Include them.
• Treating CES as a target rather than a diagnostic — if agents are incentivized on CES, they will close tickets prematurely or coach customers toward favorable responses.
• Ignoring written feedback — the comment field of a CES survey is the most useful qualitative source in the entire support stack. Customers who rate 3-5 explain exactly which step was the friction point.