Where CSAT asks whether the customer liked the outcome, CES asks whether the path to the outcome was painful. That shift from feeling to friction is why CES catches churn risk other scores miss.
CES, or Customer Effort Score, measures how much effort a customer had to put in to get an issue resolved or a task completed. It is usually captured on a 1 to 5 or 1 to 7 agreement scale with a statement like "the company made it easy to handle my issue." CES is the best single predictor of future loyalty for support and service interactions, because exhausted customers quietly leave even when satisfaction scores look fine.
The six things to know before launching a CES program, and the single reason this score tends to predict churn more reliably than its more famous siblings.
CES is a short survey anchored to a statement like "the company made it easy to resolve my issue." The customer agrees or disagrees on a 1 to 5 or 1 to 7 scale. The score reports how hard the customer had to work, which is a different question than how happy they ended up.
The modern CES uses a 1 to 7 agreement scale, where 7 is strongly agree and 1 is strongly disagree. A 1 to 5 version is common on mobile. Report the top-two-box percentage, not the mean, so the number reflects the share of customers who found the path genuinely easy.
Fire the CES survey immediately after the moment of effort: a support ticket closes, a self-service article is used, an onboarding step completes, a renewal is signed. The signal degrades fast once the friction fades from memory, so cadence beats volume every time.
Research first published by CEB in the Harvard Business Review found that effort is a stronger predictor of future loyalty than satisfaction or delight in support contexts. Customers forgive missed targets. They do not forgive making them work for a resolution they expected to be easy.
A customer can rate CSAT high because the resolution worked, then rate CES low because getting there took three calls and two escalations. That gap is the churn risk other scores miss. CES surfaces the exhausted segment that renews once and leaves quietly after.
CES is a service and support metric. Using it to measure a demo call or a brand campaign produces nonsense, because effort was not the point of that interaction. Pair the right score with the right moment, or the dashboard looks busy and signals nothing.
The CES formula is deliberately simple. The interesting choices are around scale, question phrasing, and reporting style. Teams routinely report CES as a mean, which smears the signal across neutral responses. The six cards below are the standard version and the common distortions worth avoiding before the first survey ships.
CES percent equals the number of responses that agreed the experience was easy, divided by total responses, multiplied by 100. On a 1 to 7 scale, agreed usually means a 5, 6, or 7. On a 1 to 5 scale, agreed usually means a 4 or 5. The result is a clean percentage ready to trend.
The modern CES is phrased as a statement: "The company made it easy for me to handle my issue." The customer picks a point on the agreement scale. That phrasing anchors the respondent on effort, not emotion, which is what separates CES from CSAT when both run on the same ticket.
CEB's 2010 research introduced a 1 to 5 scale. The updated 2013 version switched to 1 to 7 for better resolution between neutral and strong responses. Most mature programs now run 1 to 7. Keep whichever scale you pick stable for at least four quarters, because changing scales breaks the trend line.
Reporting the mean hides the shape of the distribution. A dataset with half sevens and half ones averages to a four, which looks neutral and is actually two very different experiences. Report the share of agreement responses so a bimodal pattern stays visible on the dashboard.
If 100 tickets close and 15 customers respond, CES is calculated on 15, not 100. A high score on a 10 percent response rate usually means the frustrated customers never answered. Track the response rate alongside the score, and segment-flag anything below 20 percent as suspect data.
Pair the scale with a single open-text field asking "what made this harder than it should have been?" The qualitative replies are the specific friction, which is what feeds coaching, policy changes, and product fixes. The number alone is a trend. Number plus comments is a learning system.
Teams routinely conflate CES, NPS, and CSAT, then wonder why the dashboards disagree. Each score answers a different question at a different cadence. The six cards below are the shortest honest comparison. Pick the right score for the moment and the three stop fighting each other.
Statement: the company made it easy. 1 to 7 agreement scale. Fired after a resolution or a self-service attempt. Reports the share of customers who found the experience easy. Best for support, service, onboarding, and renewal interactions where effort is the main variable the customer will remember.
Question: how satisfied were you with this interaction. 1 to 5 or 1 to 7 scale. Fired at a touchpoint, right after it closes. Reports the share of satisfied responses. Best for coaching individual reps, tuning a specific workflow, and spotting a bad week before it compounds into a quarter.
Question: how likely are you to recommend us. 0 to 10 scale. Calculated as promoters minus detractors. Fired quarterly or at a milestone, not after every ticket. Best for board-level trending, cross-segment comparison, and tying experience to growth over longer time horizons.
CES predicts whether the customer is quietly losing patience. CSAT predicts whether the next ticket will be easier or harder. NPS predicts whether the account will renew and whether it will refer new business. Three leading indicators, three timescales, three audiences. Running all three is normal.
CES fires after every service resolution. CSAT fires after every closed touchpoint. NPS fires on a quarterly rotation, so the sample stays comparable and the customer is not surveyed into fatigue. Three cadences, three audiences. Run them in parallel on different trigger rules.
Use CES when a workflow should feel effortless: a password reset, a billing correction, a renewal, a self-service article, an onboarding step. Use CSAT when a touchpoint should feel good. Use NPS when a relationship should feel worth keeping. The right score at the right moment is the whole trick.
CES is a service and support metric. It is the right score when a customer expected an interaction to be easy and the thing to measure is whether the path lived up to that expectation. The six cards below are the moments where CES tends to outperform CSAT or NPS as a leading indicator, and where to leave CES out of the survey plan entirely.
The classic CES moment. A customer opened a ticket to fix something. The ticket closes. The CES survey fires. The score reports whether the fix took one touch or five, whether the customer had to repeat context, whether the agent had the authority to resolve. The strongest predictor of whether the next ticket is coming.
The customer tried to help themselves. The article closed, the bot ended, the knowledge base search ran. CES captures whether the attempt actually worked or whether the customer gave up and will escalate next. Deflection is only a win if the self-service path was easy, not just available.
A new customer finishes activation, imports data, invites teammates, hits first value. CES reports whether the path to that milestone felt engineered or improvised. Onboarding effort predicts expansion revenue months later, because the customers who found setup easy are the ones who keep adopting new workflows.
The renewal closed. CES reports whether contracting, procurement, legal review, and finance felt like a routine or a slog. High-effort renewals quietly become churned renewals twelve months later, because procurement remembers friction longer than the champion does. Measure it the moment the signature lands.
A migration, an integration project, a custom build. CES reports whether the engagement felt like a partnership or a procurement exercise. Services effort scores are a near-direct predictor of whether the customer buys the next engagement from the same team or shops it to a competitor.
CES is not for pre-sales demos, marketing campaigns, product launches, or community events. Effort was not the main variable the customer will remember from those moments. Use CSAT for moments and NPS for relationships. Running CES on everything dilutes the signal and burns through response rates.
A CES program lives or dies on the wording of the one question. Primed phrasing produces inflated scores. Vague phrasing produces noisy ones. The six cards below are the design decisions that separate a CES signal that drives change from one that looks fine on a slide and does nothing.
The modern CES is phrased as a statement the customer agrees or disagrees with: "The company made it easy for me to handle my issue." The agreement format anchors the respondent on effort instead of mood, and it reads consistently across cultures and languages where question phrasing drifts.
Avoid wording like "our excellent team made it easy" or "despite the difficulty, how easy was it." Both prime a direction. Keep the sentence neutral and short. If the customer has to parse the question, they are no longer answering it, they are decoding it, and the score becomes noise.
Replace "handle my issue" with the actual task: "resolve my billing question," "complete my onboarding," "sign my renewal." The specific noun tells the customer which effort you are asking about, which keeps the score tied to the touchpoint instead of drifting toward the overall relationship.
Label all seven points on a 7-point scale (strongly disagree through strongly agree) rather than only the endpoints. Fully labeled scales reduce respondent guessing, improve consistency between respondents, and make the distribution easier to interpret when you look at the raw data later.
Most CES responses come from a phone, often in a notification tap seconds after the ticket closes. A 7-point scale has to fit on a mobile screen without horizontal scroll. Test the survey on the smallest phone your audience uses before shipping. A survey that looks fine on desktop and squeezes on mobile just loses response rate.
Pair the scale with one optional open-text field: "what made this harder than it should have been?" The replies are the specific friction, named by the customer, which is exactly what coaching, process redesign, and product fixes need to work from. The number alone cannot do that job.
Benchmarks are useful as rough calibration and misleading as a target. Scoring 75 is only impressive if last quarter you were at 68, and the scale, phrasing, and response rate all have to match before any two numbers can be compared honestly. The six cards below are the ranges most practitioners cite, with the usual caveats about method consistency.
Business software support hovers in the high 60s to high 70s on top-box agreement. Onboarding tickets usually score lower than steady-state ones, because product knowledge is still forming. Watch the gap between first-month and month-six scores: a wide gap points to a product-education problem, not a support-quality one.
Shipping, returns, and refund flows skew high when the policy is generous and skew sharply when it is not. Order-related tickets and account-related tickets diverge enough that reporting one average blurs more than it reveals. Segment by ticket type before any benchmark comparison is honest.
Banking, insurance, and fintech score lower than other industries, because the outcome a customer wants (reversing a fee, approving a claim) is often not the outcome the policy allows. The useful signal is the trend inside the "policy denied" segment, not the headline percentage.
Telecom and utility support historically score among the lowest CES ranges, largely because the ticket mix includes a lot of billing disputes, service outages, and account-change procedures with long resolution chains. Trend improvement here moves retention faster than almost any other industry.
CES on help articles, chatbots, and knowledge-base deflection tends to run below human-support CES, because the gap between what self-service covers and what the customer actually needed is often wider than the design intended. Low self-service CES is usually a content gap, not a UX one.
Industry benchmarks make a useful sanity check and a bad target. Scale, phrasing, response rate, segment mix, and sample size vary across every published number. The comparison that actually matters is this quarter versus last quarter, holding the method constant across both.
A CES program that lives in a separate survey tool is a CES program that stays disconnected from the record it is measuring. When the score lands on the customer timeline next to the ticket, the call, and the renewal, the action loop becomes a routine instead of a project. The cards below are the shape CES takes when it is run from inside the CRM instead of alongside it.
A ticket closes, a self-service article is marked helpful, a renewal signs, an onboarding step completes. The CRM fires the CES survey the moment the record flips, not on a nightly batch. Scores land on the customer timeline next to the interaction that produced them, so context never has to be reconstructed later.
Support can run a 1 to 7 agreement scale with the standard statement. Onboarding can run a 1 to 5 with task-specific wording. Renewals can run a two-question variant. Each touchpoint type has its own survey template, question, and low-score threshold. One tool, every moment, no end-of-quarter reconciliation.
A score below the configured threshold creates a task on the account owner or routes to the on-duty CSM, with the ticket, the open-text comment, and the full customer context attached. SLA clock starts. The callback is a workflow step, not a reminder in someone's head. The loop closes the same day.
CES trends by rep, by team, by touchpoint, by segment, visible in the same reporting workspace as pipeline, renewals, and churn forecasts. Effort friction and retention live together, because they are the same conversation on different timescales. One login, one data model, one quarterly review.
Open-text comments feed a themes report so recurring friction surfaces without a human reading every reply. Themes become tasks on the right team: billing effort to finance, workflow effort to operations, product effort to engineering. The score becomes a routing engine, not just a dashboard metric.
CES inside the CRM stops being a slide and starts being a workflow trigger. Every high-effort response has an owner, an SLA, and a next step. Every trend becomes a coaching conversation. Every pattern becomes a roadmap item. The score earns its weekly review slot because it moves the work.
Strkr fires CES on every closed ticket, self-service session, onboarding step, and renewal, routes high-effort responses to the right CSM with an SLA, and surfaces trends next to pipeline, renewal, and churn reporting. One tool, one data model, every service touchpoint. Pricing is published and the platform tour is open.
CES stands for Customer Effort Score. The acronym is used both as a label for the metric and as a shorthand for the survey that produces the metric. In day-to-day business conversation, "CES" almost always refers to the score itself, usually reported as a top-box agreement percentage on a 0 to 100 scale.
The CES formula is the number of responses that agreed the experience was easy, divided by total responses, multiplied by 100. On a 1 to 7 agreement scale, agreed usually means a 5, 6, or 7. On a 1 to 5 scale, agreed means a 4 or 5. The result is a percentage, reported as top-box share rather than a mean score.
A rough industry benchmark is 65 to 78 percent for SaaS support, 70 to 82 percent for e-commerce, 55 to 70 percent for financial services, and 60 to 72 percent for self-service touchpoints. These ranges vary widely by scale, response rate, and segment mix, so the more useful comparison is your own trend line quarter over quarter with the method held constant.
CSAT asks how satisfied the customer felt about the outcome. CES asks how much effort the customer had to spend to reach the outcome. A customer can rate CSAT high and CES low, meaning the resolution worked but the path was exhausting. Pairing both scores on the same ticket surfaces the "satisfied but tired" segment that often churns a quarter or two later.
NPS measures long-term loyalty across the whole relationship on a quarterly or milestone cadence. CES measures effort tied to one specific interaction, right after it happens. NPS is a lagging, strategic score. CES is a leading, operational one. Both are useful, and most mature teams run them in parallel against different triggers rather than treating one as a replacement for the other.
Use CES when effort is the main variable: support tickets, self-service deflection, onboarding steps, renewal flows, and high-touch service engagements. Use CSAT when a touchpoint should feel good. Use NPS when a relationship should feel worth keeping. Running the wrong score at the wrong moment produces dashboards that look busy and signals that disagree with each other.
Fire a CES survey immediately after each service resolution of the type you want to measure: ticket close, self-service article used, onboarding milestone, renewal signed. Throttle to one CES survey per customer per week to protect response rate. If a customer has multiple interactions in a short window, keep the survey tied to the most recent resolution and let the others pass without a survey.
The usual mistakes are primed wording that steers the answer, running CES on moments where effort was not the point, reporting the mean instead of the top-box percentage, letting the scale drift from quarter to quarter, ignoring the response-rate denominator, skipping the open-text follow-up, and leaving low scores on a dashboard with no action loop to turn them into callbacks and coaching.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
