Customer Effort Score (CES): Definition, Question and How to Calculate It

Customer Effort Score is a transactional metric for processes. Where CSAT asks how satisfied someone was and NPS asks whether they would recommend you, CES asks a more operational question: how much effort did this cost the customer?

The idea comes from researchers at CEB (now part of Gartner). Matthew Dixon, Karen Freeman and Nicholas Toman popularised it in the Harvard Business Review article “Stop Trying to Delight Your Customers” (2010). Their central argument: in customer service, loyalty is driven less by exceeding expectations and more by how little effort customers have to spend to get their problem solved. Repeating information, being transferred between departments and having to call back are the experiences that push people away.

The first version asked “How much effort did you personally have to put forth to handle your request?” on a 1–5 scale, where a high number meant more effort. Because that reversed direction confused respondents and readers of reports, a revised version, usually called CES 2.0 and described in the authors’ later book “The Effortless Experience” (2013), replaced it with an agreement statement on a 1–7 scale, where higher is better. That is the form most organisations use today.

The CES 2.0 statement is:

“[Company] made it easy for me to handle my issue.”

Customers answer on a seven-point agreement scale:

Score	Answer
7	Strongly agree
6	Agree
5	Somewhat agree
4	Neither agree nor disagree
3	Somewhat disagree
2	Disagree
1	Strongly disagree

Adapt “handle my issue” to the process you are measuring, and keep the wording fixed afterwards:

• “It was easy to get my question answered today.”
• “Returning my order was easy.”
• “Booking my appointment was easy.”
• “It was easy to complete my application at this service centre.”
• “Finding what I needed on the website was easy.”

Simpler variants exist: a 1–5 “very difficult” to “very easy” scale, or three or five faces on a kiosk. They are fine as long as you use one version consistently and do not compare results across versions.

Always add one open question, ideally only for low scores: “What made this difficult?” The answers usually point straight at the step to fix.

There are three common ways to turn CES answers into a number. Use whichever your team understands best, then stick to it.

Worked example: 120 customers answer after a support case:

Score	7	6	5	4	3	2	1
Responses	24	36	30	12	8	6	4

1. Average CES = sum of all scores ÷ number of responses. (7×24 + 6×36 + 5×30 + 4×12 + 3×8 + 2×6 + 1×4) ÷ 120 = 622 ÷ 120 = 5.18 out of 7.
2. % agree (easy) = answers 5–7 ÷ all answers × 100. (24 + 36 + 30) ÷ 120 = 90 ÷ 120 = 75%.
3. Net easy = % agree − % disagree. Answers 1–3 are 18 of 120, or 15%, so 75 − 15 = 60.

These are the default numbers in the calculator above. Change them to your own counts, or use the standalone CES calculator.

The average is the most common headline; the percentage is easier to explain to a frontline team (“three in four customers found it easy”). Whichever you report, keep the share of 1–3 answers visible, because a stable average can hide a growing group of customers who struggled.

On the 1–7 agreement scale, 4 is the neutral midpoint, so an average above 4 means more customers agreed than disagreed that the process was easy. Beyond that, there is no meaningful universal benchmark. Effort depends almost entirely on the process: resetting a password and settling an insurance claim will never score the same, and comparing them, or comparing your claims process with another company’s password reset, tells you nothing.

Useful comparisons are internal:

• The same process over time, before and after a change.
• The same process across channels: is the return easier in store or online?
• The same process across locations or teams, with enough responses per group.

Read the distribution, not just the average. A rising share of 1–3 answers is an early warning even when the average barely moves.

CES works best immediately after a task the customer had to complete:

• after a support ticket, call or chat is closed;
• after a return, exchange, refund or claim;
• after onboarding, account opening or an application at a service centre;
• after self-service actions: booking, rescheduling, paying a bill.

It is less useful for experiences that are not tasks (a hotel stay, a meal, a museum visit), where CSAT fits better, and it says little about the overall relationship, which is the job of NPS. The three metrics are complementary; the comparison NPS vs CSAT vs CES shows how to combine them.

CES is also a strong process-improvement tool because the fixes are usually concrete: fewer transfers, fewer forms, clearer instructions, a callback instead of a queue. When the open comments name the same step again and again, that step is your next project.

1. Ask within minutes or hours, not days. Effort is forgotten quickly. Trigger the survey by SMS or email when the ticket closes, or show it on a tablet at the service desk.
2. Keep it to two questions: the CES statement and “What made this difficult?” (shown for 1–3). Add one driver question only if you will act on it.
3. Measure one process per survey. “Was it easy?” only makes sense when the customer knows exactly what “it” is.
4. Segment by channel and reason. A CES for “all contacts” blends easy and hard requests; split by contact reason to see where effort really sits.
5. Act on low scores fast. A customer who strongly disagrees that it was easy may not have had the issue solved at all. Route those answers to a team member for a follow-up, as in closed-loop feedback.

The business case is the same as for any loyalty metric: PwC’s 2018 “Experience is Everything” research found that almost a third of customers (32%) will walk away from a brand they love after one bad experience, and needless effort is one of the most common ways an experience turns bad.

Qmeter does not lock you into one CES format. You build the CES question in the survey builder as a 1–7 single-choice question (or a smiley rating on kiosks), add a conditional “What made this difficult?” question for low scores, and send it through any channel:

• SMS and email campaigns after a case or appointment, QR codes at service points, kiosks and tablets at the desk (online or offline) and web links or widgets after self-service tasks.
• Ticket generation rules that turn low ratings into follow-up cases with an owner and SLA.
• Reports and AI analysis of open comments, so you can see which steps generate the most effort, by location, service and channel.

Qmeter’s own tickets also send an automatic satisfaction survey after resolution, so you can see how customers rate the way complaints were handled. Start a 14-day free trial, or read about ticketing and SLAs.