--- path: /blog/csat-calculator title: "CSAT Calculator: Work Out Your Customer Satisfaction Score" description: "Free CSAT calculator: work out your customer satisfaction score with the top-2-box method, see an honest margin of error, and compare 2026 benchmarks." canonical: https://www.shine.studio/blog/csat-calculator author: "Travis Keeney" publishedAt: 2026-07-16 topic: "Best Practices" --- # CSAT Calculator: Work Out Your Customer Satisfaction Score CSAT is calculated with the top-2-box method: divide the number of respondents who rated you 4 or 5 on a five-point satisfaction scale by everyone who answered, then multiply by 100. The free customer satisfaction score calculator below does the math, shows how much to trust the result, and compares it against current industry benchmarks. It loads with a worked example; replace the numbers with your own.
Interactive CSAT calculator: enter your satisfied and unsatisfied counts (or a full 1 to 5 rating tally) to get your score, Wilson confidence interval, and an industry benchmark comparison. Use it in a browser at https://www.shine.studio/blog/csat-calculator
Everything runs in your browser. Nothing you type is sent anywhere; the share link just encodes your counts in the URL, and the downloadable score card is drawn on your device. ## How is CSAT calculated? To calculate CSAT, count the satisfied responses and divide by the total. The convention, as Qualtrics documents it, is to use "the responses of 4 (satisfied) and 5 (very satisfied)" on a 1 to 5 scale: | Rating | Counts as | |---|---| | 5 (very satisfied) | Satisfied | | 4 (satisfied) | Satisfied | | 3 (neutral) | Not satisfied | | 1 or 2 | Not satisfied | Worked example: you survey a support queue and collect 250 responses. 210 people rate the interaction 4 or 5. Your CSAT is 210 divided by 250, times 100: **84%**. The detail people miss is the 3s. A neutral rating feels harmless, but the top-2-box method counts it against you, exactly like a 1. That is deliberate: a customer who cannot say they were satisfied was not satisfied, and folding neutrals into the win column is how teams end up celebrating a number their churn rate disagrees with. ## What is a good CSAT score? The commonly cited range comes from Zendesk's guidance: "Many industries consider a CSAT score between 75 and 85 percent as good and a score above 90 percent as exemplary." | CSAT range | Interpretation | |---|---| | Below 75% | Below the good range | | 75% to 85% | Good | | 85% to 90% | Between good and exemplary | | Above 90% | Exemplary | The calculator above folds the middle two rows into one "Good" band for its readout; the boundaries that matter are 75 and 90. Retently's 2026 analysis draws softer lines from its survey data: 65% to 80% is "a dominant value across industries," and 70 to 90 is "a good place to be in." As with every satisfaction metric, the range that matters most is your own: the same 78% is a slump for a consulting firm and a standout for a healthcare provider, as the benchmarks below show. ## CSAT benchmarks by industry (2026) Retently's 2026 CSAT benchmarks (published April 2026) use the same top-2-box method as the calculator above, so the comparison is like for like: | Industry | Average CSAT (2026) | |---|---| | Consulting | 83% | | Digital Marketing Agencies | 83% | | Financial Services | 81% | | Property Management | 78% | | Ecommerce & Retail | 77% | | Education | 68% | | Internet & Software Services | 64% | | Healthcare | 57% | | Construction | 30% | | Communication & Media | 26% | Read the bottom of that table with care: scores like 26% say as much about how and when those industries survey as about how customers feel. Benchmarks vary by who measures and how.
You will also see the American Customer Satisfaction Index cited as a benchmark: the national ACSI stood at 76.7 in Q1 2026. Be careful comparing against it: ACSI is an econometric index on a 0 to 100 scale, not a top-2-box percentage, so an ACSI of 77 and a CSAT of 77% are different animals. Compare your top-2-box score against top-2-box benchmarks only.
## How many responses do you need for a reliable CSAT? CSAT is a single proportion, and small samples flatter it. Five happy customers in a row reads as 100%, but the honest math says the true rate could plausibly be anywhere above 57%. That is why the calculator above uses the Wilson score interval, one of the two methods Sauro and Lewis's research on small-sample confidence intervals found gave coverage closest to the nominal 95%. MeasuringU's guidance is blunt that the naive textbook interval "should be avoided" below about 100 responses; the corrected interval it recommends outright, which "should be used almost all the time," is the adjusted Wald, Wilson's near-twin built on the same recentering. The practical consequences, for a score around 80%: at 25 responses the calculator above reports a margin of error of ±19 points, at 100 it is ±9, and it takes about 400 responses to get down to ±4. (Those figures are the interval's largest distance from your score; the asymmetric range itself is always shown alongside.) The margin shrinks with the square root of the sample, which is why it takes four times the responses to cut it in half. A week-over-week CSAT wobble smaller than your margin of error is noise, not news. Volume is easier to get with CSAT than with other survey metrics. Retently's 2026 response-rate study (25 million invitations across 600 ecommerce brands) measured CSAT response rates at 9.76%, roughly double the 4.5% that NPS surveys managed, because CSAT arrives right after an interaction the customer actually remembers. ## CSAT vs NPS: which should you use? CSAT grades a moment; NPS grades the relationship. Qualtrics puts the distinction plainly: "one is more transactional (CSAT), and one is more relational (NPS)." CSAT tells you the checkout flow annoyed people this week; NPS tells you whether customers would stake their reputation on recommending you at all. Most programs run both: CSAT wired to touchpoints (support tickets, onboarding, delivery) to localize friction, and NPS on a slower cadence to watch the relationship. If you are doing the NPS math too, the sibling of this tool, our free NPS calculator, works the same way, down to the honest margin of error. ## How do you improve your CSAT? Fix the touchpoint, not the survey. CSAT is transactional, so the verbatims tell you exactly which moment failed: read every 1 and 2, close the loop with those customers directly, and look for the repeat offender in the comments. The 3s deserve special attention because they are the cheapest wins on the board: a neutral customer was usually one concrete fix away from a 4, and the top-2-box math means converting one is worth exactly as much as converting an angry detractor. Then do something with the 5s. A customer who just rated an experience "very satisfied" is at peak goodwill, and in most companies that moment scrolls away unused. Asking for a review, a referral, or a short interview while the experience is fresh converts the rating into an asset you can put in front of the next buyer. ## Where Shine fits (and where it doesn't) Shine runs CSAT and NPS surveys free, over email and an embeddable website widget, with no cap on responses. The product's actual job begins after the score: when a 5 comes in, Shine invites that customer into a short AI-conducted interview and turns the conversation into verified quotes, case studies, and review drafts. If your bottleneck is what to do with satisfied customers once you find them, that is what it is built for. To be straight about the boundaries: if you need enterprise survey analytics across millions of responses with complex routing, a Qualtrics-class platform is the right category. And if all you need today is the math, the calculator above requires no signup and never will. ## Frequently Asked Questions **What scale should a CSAT survey use?** The 1 to 5 scale is the convention the top-2-box method assumes, and it is what most benchmarks (including the Retently figures above) are built on. Teams sometimes use 1 to 3, 1 to 7, or 1 to 10 scales; those can work internally, but every scale change redefines what "satisfied" means, breaks comparability with published benchmarks, and resets your own trend line. Pick 1 to 5 unless you have a strong reason not to, and never change scales mid-stream. **Is a 3 counted as satisfied?** No. Top-2-box counts only 4s and 5s as satisfied; a 3 lands in the same bucket as a 1. Some teams also track "top-box" (5s only) alongside top-2-box when they want an even stricter read, but never report it as CSAT: every published benchmark, including the ones above, assumes top-2-box, so mixing the two quietly inflates or deflates comparisons. **Why is my CSAT high while customers still churn?** Three usual suspects. CSAT is transactional, so it can be genuinely high at the touchpoints you measure while the overall relationship decays at ones you don't. Response bias inflates it: satisfied customers answer more readily than quietly frustrated ones. And a high score built on a handful of responses may just be noise; check the margin of error before celebrating. Relationship-level questions are NPS territory, which is exactly why programs run both. **How often should you send CSAT surveys?** Trigger them on events, not calendars: after a support resolution, a delivery, onboarding completion. The rhythm belongs to the interaction. Cap how often any one customer can be asked (once every few weeks is a common ceiling) so frequent users don't get carpet-bombed, and keep the survey to one question plus an optional comment; every extra field costs responses.
Don't want to do this yourself? Shine runs CSAT for you, free: the survey goes out after the moments you care about, the math happens automatically, and the 5s get turned into proof. The calculator above stays free and signup-less either way.