Survey-based CES misses the customers who worked hardest to get help. Here is how to measure customer effort from the feedback and support conversations you already collect.
Customer Effort Score (CES) was built as a survey question, but the highest-effort customers rarely answer it. You can measure effort continuously from open-ended feedback and support conversations by reading signals like repeat contacts, escalations, and channel switching. Thematic's Scoring Agent produces a predicted effort score from unstructured feedback and traces every point back to the themes and verbatims driving it.
Most teams measure customer effort by adding one more survey question: "How easy was it to resolve your issue?" The problem is that the customers who worked hardest to get an answer are the least likely to stop and rate you. Effort is the experience that quietly drives people away, and the survey that is supposed to catch it misses the people it most needs to hear from.
You can measure customer effort without waiting on a survey. Thematic reads effort directly from the feedback and support conversations you already collect: support tickets, call summaries, chat transcripts, app store reviews, and open-ended survey comments. It turns those signals into a trackable effort score. Each score traces back to the exact themes and verbatims behind it. The survey question becomes optional. The behavior in your existing data does the measuring.
This article covers what customer effort and the Customer Effort Score (CES) actually mean, why the survey version misses so much, which signals in open-ended feedback and support conversations reveal effort, and how Thematic scores effort continuously and ties every point of movement back to a cause.
Customer effort is the amount of work a customer has to do to get their problem solved or their goal met. The Customer Effort Score (CES) is the metric built to capture it. CES was created by CEB (now part of Gartner) and introduced in the 2010 Harvard Business Review article "Stop Trying to Delight Your Customers," based on a study of more than 75,000 customer interactions.
CES shows up in two standard forms:
Effort matters because it predicts loyalty better than satisfaction does. Gartner found that customer effort is 40% more accurate at predicting loyalty than customer satisfaction. The CEB research behind CES found that 96% of customers who have a high-effort interaction become more disloyal, compared with just 9% of those who have a low-effort one. Low effort isn't a nice-to-have. It's the single clearest signal of whether a customer stays.
Deriving a defensible effort measure from open-ended feedback and support conversations, rather than a survey question, takes four things.
A clear definition of effort in your own words. Before anything is scored, you need to state what "high effort" looks like for your business: repeated contacts, escalations, having to repeat information, or being bounced between channels. The definition drives everything downstream.
Coverage of the conversations where effort actually shows up. Effort is most visible in support tickets, call and chat transcripts, and agent notes, not in a survey field. The measure has to read those unsolicited sources, not just solicited ones.
Bottom-up theme discovery. Effort has causes. A score that says effort rose is only useful if it also surfaces the themes driving it, such as a broken password reset or a confusing returns policy.
Traceability back to the raw comment. Any effort number a skeptical leader will act on has to trace back to the specific verbatims that produced it. A black-box score invites the argument instead of ending it.
A CES survey is a useful benchmark, but as the primary way to measure effort it has structural gaps:
A useful demo-time test: ask whether a tool can tell you not just that effort went up this quarter, but which theme drove the change and which customer comments prove it. If the answer is only a number, effort is being measured too shallowly to act on.
Thematic measures effort by reading the feedback and conversations customers generate anyway, then scoring effort from what they actually said and did. There are three parts.
Thematic reads the signals that reveal effort. Effort leaves fingerprints in unstructured data: customers who contact you more than once about the same issue, tickets that escalate, sessions that switch channels, and language like "still not resolved" or "third time asking." Thematic detects these behavioral and linguistic signals across support tickets, call summaries, chat, reviews, and open-ended survey comments in one place.
Thematic's Scoring Agent turns those signals into a predicted effort score. Rather than requiring a survey, the Scoring Agent generates a predicted effort score directly from unstructured feedback, alongside metrics like predicted NPS and churn propensity. You define the metric in plain language and validate it with a human in the loop, so the score reflects how your business defines effort.
Every score traces back to its drivers. Each effort score in Thematic links to the themes and the verbatims that produced it. When effort moves, you can go from "effort rose three points this month" to the specific theme behind it and the exact comments customers left. Scores quantify the outcome; themes explain the driver.
Because effort is measured continuously from data you already own, the measure updates as issues emerge, instead of once a quarter when the survey closes.
Atom Bank cut calls on its highest-effort issues. Atom Bank, a UK digital bank and a Thematic customer for more than three years, unified feedback from app store reviews, Trustpilot, support-center complaints, and Salesforce call summaries across three product lines and seven channels. By finding the highest-effort contact reasons in that feedback and fixing them, Atom Bank saw a 69% drop in calls about unaccepted mortgage requests, a 43% drop in calls about savings maturities, and a 40% drop in device-related calls. Contact-center failure demand fell 30%. Every one of those reductions is effort removed from a customer's day, measured from feedback rather than a CES survey.
Vodafone New Zealand aligned its teams around what the feedback said. The telco used Thematic to identify themes in NPS verbatims and route them to the right teams, saving 60 hours every month of manual reading and posting a double-digit increase in touchpoint NPS within nine months. "Thematic helps us identify themes in customer feedback which informs where our teams should focus their attention," said Tania Parangi, NPS Evolution Manager at Vodafone New Zealand. "They also clearly show us the positive and negative impact of the changes we're making, so we can adapt and refine in real time."
A New Zealand water utility took the same approach to complaints after major storms, using AI analysis of support-center feedback to surface its biggest sources of effort, long wait times and communication gaps, then reprioritizing proactive communication to cut repeat contact.
Before choosing how to measure customer effort, ask:
You measure customer effort from open-ended feedback and support conversations by reading the behavioral and linguistic signals of effort, such as repeat contacts, escalations, and friction language, and scoring them with AI text analytics instead of relying on a survey question. Thematic does this with its Scoring Agent, producing a predicted effort score that traces back to the themes and verbatims driving it. To test any tool, ask it to show you not just that effort changed, but which theme moved it and which comments prove it.
Thematic turns fragmented feedback into one consistent source of customer truth — so every team acts on the same customer story. Up and running in days, not quarters.
Transforming customer feedback with AI holds immense potential, but many organizations stumble into unexpected challenges.
