Customer Satisfaction Survey Template
Most businesses know whether customers are happy in the way most people know whether they’re healthy — by vibes, anecdotes, and the loudest voice in the room. The CSAT discipline replaces that with something colder and more useful: a single satisfaction question, asked consistently across your base, producing a number you can trend, segment, and defend in a meeting. This template builds that instrument.
The Metric, Defined Before It’s Measured
CSAT only works as a managed number if its definition never wobbles:
- The question — how satisfied are you with [your product / our service / your experience]? on a fixed rating scale. Pick the wording, freeze it, and never redesign it mid-year — a reworded question breaks every comparison you’ve built
- The population — who gets asked, and how they’re chosen. A score from whoever happened to email you is a complaint count; a score from a proper sample is a metric
- The calculation — typically the share of top-box responses (4s and 5s on a 5-point scale). Decide once, calculate it the same way every cycle, and write the method down
Sampling: The Part Everyone Gets Wrong
The difference between a metric and a mood is sample discipline:
- Survey a defined sample each period, not the whole base at once — rotating cohorts keep customers fresh and responses honest
- Size the sample before collecting it. Work out how many responses you need for the confidence you want; thirty responses from a ten-thousand-customer base is an anecdote wearing a percentage
- Watch for responder bias — the delighted and the furious over-respond; the content middle stays home. A consistent sample method is the antidote, because the bias at least stays constant across cycles
What the Trend Tells You
One CSAT reading is trivia; the line across quarters is management information:
- Drift is the early warning. Satisfaction rarely crashes — it slides, two points a quarter, invisibly, until a renewal quarter makes it expensive. A tracked trend catches the slide while it’s still a slide
- Segment the score. By plan, by cohort, by product line — export and filter responses to find where the average is hiding a problem. A healthy 82% can contain an onboarding cohort sitting at 61%
- Pair every drop with a why. Keep one optional open field beside the score; when the line moves, the comments explain the movement
The Review That Makes It Management
A metric nobody reviews is decoration. Put the number on a rhythm: each cycle, score it, compare against baseline, segment for outliers, and attach one page of commentary — what moved, why, what changes. Fifteen minutes per cycle, and the business gains the thing vibes can never provide: a satisfaction number that means the same thing in every meeting it appears in.
Frequently Asked Questions
What is a CSAT score, precisely?
A customer satisfaction score: responses to a standardized satisfaction question, aggregated — usually as the percentage of top-scale responses. It answers one question narrow enough to be honest — are customers satisfied? — and its power comes from trending that answer over consistent cycles.
How is CSAT different from NPS?
Different questions, different jobs. CSAT measures satisfaction with what you deliver; NPS measures willingness to recommend, which blends satisfaction with loyalty and brand feeling. They move together loosely and diverge instructively — a high CSAT with a low NPS means customers are served but not attached. Run CSAT as your operational metric and NPS as your relationship metric rather than forcing one to do both.
How many responses do we need before trusting the number?
Enough that one grumpy respondent can’t move it — typically a few hundred for a mid-size base, calculated properly against your population size and desired confidence. Under-sample and you’ll read noise as trend; the sample-size calculation takes two minutes and prevents a year of arguments about whether the score really dropped.
What CSAT score should we aim for?
Set the target against your own baseline, not folklore. If your first honest measurement lands at 76%, the goal is 80 — not an industry number of contested origin. The trend and the segments carry the insight; the absolute value mostly flatters or alarms people who haven’t read their own history.
Should the satisfaction question include an open comment field?
One optional field, yes — never required. The score tells you that satisfaction moved; the comments tell you why, and the why is where management action lives. Requiring explanations suppresses responses from exactly the content middle your sample most needs.