Customer satisfaction score: measuring CSAT so it changes something

A customer satisfaction score — CSAT — asks how satisfied someone was with a specific interaction, usually on a scale of one to five, and reports the percentage who answered in the top one or two boxes. Unlike NPS, which asks about the relationship as a whole, CSAT is deliberately narrow: it measures the support ticket, the delivery, the onboarding call, the thing that just happened.

That narrowness is its strength. A relationship score tells you the temperature; a transactional score tells you which process is failing. If you can only run one, run CSAT on the interaction that matters most to your customers and act on the results within the week.

Designing the question

• Ask about one interaction, named specifically: "How satisfied were you with how we handled your request?" rather than "How satisfied are you with us?"
• Use a consistent scale — five points is standard, with labels rather than bare numbers so everyone interprets them the same way.
• Define the calculation before you start: the percentage answering 4 or 5 out of the total who answered. Write it down, because changing it later invalidates your history.
• One follow-up: what would have made it better. Optional, free text, and the only field that tells you what to change.
• Send it immediately after the interaction, while the detail is fresh. CSAT sent a week later measures memory.
• Attach context automatically — which agent, which product, which channel — instead of asking.
• Keep it to one screen and no login. Response rate collapses at the first obstacle.

Watch the response rate as carefully as the score. CSAT is systematically biased towards people with strong feelings, and if only 8% respond, a comfortable 92% score may be describing a small, self-selected group. A rising score alongside a falling response rate is a warning, not an achievement.

Turning scores into changes

1. Choose the one or two interactions where satisfaction genuinely affects whether people stay. Measuring everything produces data nobody reviews.
2. Set the trigger so the survey goes out automatically after the interaction rather than when someone remembers.
3. Route every low score to a person the same day. Speed matters more than the eventual response.
4. Tag the comments into themes and count them weekly. Themes with counts are actionable; individual quotes are anecdotes.
5. Segment by agent, channel and product before drawing conclusions. An average across all three usually hides the actual problem.
6. Pick one theme per month and fix it, then check whether that theme's frequency falls.
7. Report the score with the response rate and the sample size beside it, always.
8. Review the question annually — and if you change it, keep the old series separate rather than pretending the numbers are comparable.

CSAT, NPS and CES

Three metrics get confused constantly. CSAT measures satisfaction with a specific interaction. NPS measures willingness to recommend the relationship. CES — customer effort score — asks how easy it was to get something done, and is the best predictor of the three for whether a service interaction damaged loyalty. Running all three at once is a common mistake for a small company: it triples the survey burden on customers and produces three numbers nobody has time to act on. Pick the one that matches the decision you want to make.

Ettex Sheets is where the measurement lives: import an existing XLSX with formulas intact, one row per period with responses, top-box count, score and response rate calculated by formula, conditional formatting to flag a week where the score or the response rate drops, charts for the trend, pivots to break the score down by agent, channel or product, and cell comments so an unusual week carries its explanation. Version history keeps the series honest when a definition changes. The survey itself is a form in Ettex Forms, landing timestamped in a shared inbox with export to CSV or XLS; the invitation goes out through Ettex Mail; and the customer context sits in Ettex CRM.

Said plainly: Ettex does not measure customer satisfaction for you. There is no CSAT feature, no automatic triggering after a support interaction, no score calculation, no dashboards, no alerting on low scores and no text analysis of comments. You build the form, export the responses and calculate the score in a sheet. Proportionate for tens or hundreds of responses a month; beyond that, use a product built for it.

One question asked consistently travels further than a long questionnaire. An NPS survey asks how likely someone is to recommend you on a scale of nought to ten, then reports the balance of promoters against detractors.

Why CSAT gets ignored

• Measured on everything, producing data nobody has time to review.
• Sent days after the interaction, when the detail has faded.
• Reported without the response rate, hiding severe self-selection.
• Low scores not followed up, which teaches customers that answering is pointless.
• Averaged across agents, channels and products, hiding the one that is failing.
• The calculation changed mid-series without keeping the old one separate.
• Used as an individual performance target, after which agents start asking customers for good scores and the data stops being real.
• Comments collected and never themed, so the actionable half of the survey is discarded.

A customer satisfaction score is useful when it covers one named interaction, arrives immediately, is reported with its response rate, and produces a same-day call on every low answer.