CSAT Score Calculation: Formula, Example, and Tips

CSAT score calculation is the process of turning customer satisfaction survey answers into a consistent measure. A common method divides positive answers by valid completed ratings and multiplies by 100. On a five-point scale, teams often count ratings of 4 or 5 as positive. The result is a percentage; an average rating out of five is a separate measure.

What the CSAT score measures—and what it does not

CSAT measures satisfaction with a particular product, service, or experience. In support, that might be a customer’s view of a recent interaction; for a store, it might be a purchase experience. It is a snapshot of a specific moment, not a complete measure of loyalty or the whole customer relationship.

A CSAT survey usually asks how satisfied a person was with an experience, product, or service. The question should make clear what the respondent is rating. “How satisfied were you with the help you received today?” measures a support interaction. “How satisfied are you with your order?” may capture more than the support exchange.

That distinction matters when you interpret the result. A low score after a support conversation could reflect the answer, the time taken to resolve the issue, or an underlying problem such as a delayed order. The score tells you that some respondents were not satisfied with the measured experience. It does not, by itself, explain why.

There are two common ways to report numeric ratings. A percentage CSAT counts positive responses and divides them by valid completed ratings. An average rating adds the numeric answers and divides by the number of respondents. These measures answer different questions, so label them clearly rather than treating them as interchangeable.

CSAT can help teams spot friction in a particular part of the customer journey. It does not capture every aspect of a relationship, and it should not stand alone as a measure of long-term customer behavior. Pair it with comments and operational context to investigate what may be driving the result.

How to calculate CSAT: the standard formula and positive five-point responses

For percentage CSAT, divide the number of positive responses by the number of valid completed ratings, then multiply by 100. On a five-point satisfaction scale, a common convention is to count ratings of 4 and 5 as positive. State that cutoff and use it consistently so comparisons mean the same thing.

CSAT (%) = (positive responses ÷ valid completed responses) × 100

A five-point question might use a scale from 1, “very dissatisfied,” to 5, “very satisfied.” With the common positive-response rule, ratings of 4 and 5 count in the numerator. The denominator is the number of valid completed ratings for that question.

The cutoff is part of the method, not a detail to leave implicit. If one report counts only the top rating while another counts the top two, the percentages cannot be compared as though they were calculated the same way. Record the question, scale labels, positive cutoff, and treatment of incomplete answers alongside the result.

You can use other scales or response formats, but keep the scoring definition visible. If you change from a five-point scale to another scale, or switch to a thumbs-up question, do not assume the resulting percentage is directly comparable to the old one. Explain what a positive response means in each report.

For a step-by-step companion to this method, see how to calculate CSAT. If your team uses different scales or wants to check an example, the CSAT score calculator guide covers related calculation approaches.

Worked example: calculate a percentage and a score out of 5

Suppose a five-point survey has 100 valid completed ratings. If 60 respondents selected 4 or 5, the percentage CSAT is 60 divided by 100, multiplied by 100, or 60%. If 20 ratings total 80 points, the average is 4 out of 5. Report these as separate measures.

For the percentage calculation:

• Valid completed ratings: 100
• Positive ratings of 4 or 5: 60
• Calculation: (60 ÷ 100) × 100
• Percentage CSAT: 60%

The result says that 60% of the valid completed ratings met the chosen positive-response rule. It does not mean that the average rating was 60% of the way to perfect satisfaction. It also does not reveal what the remaining respondents thought; inspect the distribution and comments for that context.

Now consider a separate group of 20 respondents whose ratings add up to 80 points. The average rating is 80 divided by 20, which equals 4 out of 5. That is an average, not 80% CSAT. A percentage conversion from an average can use different methods and produce different figures, so do not quietly substitute one for the other.

When sharing results, write the measure in full: for example, “percentage of respondents rating the interaction 4 or 5” or “average rating out of 5.” That small labeling habit helps prevent a team from comparing unlike numbers in a weekly report or a support review.

How to handle neutral, skipped, and incomplete answers

Choose a clear rule for which answers count as valid ratings before calculating CSAT. Under a common five-point setup, a neutral answer stays in the denominator but does not count as positive. A skipped answer is not a rating, so exclude it from the rating denominator and track survey participation separately.

On a five-point scale, the neutral response is typically the middle rating. If the positive cutoff is 4 or 5, a neutral rating is not positive. It still belongs in the denominator when it is a valid answer to the satisfaction question. This way, the percentage reflects all valid ratings, not just favorable ones.

A skipped question is different: the customer has not provided a rating to score. Exclude it from the denominator for that question. If someone starts a survey but skips the rating, count the survey start separately from the valid completed ratings so the reporting does not confuse engagement with satisfaction.

Survey and reporting systems may define a “response” differently. Decide what counts as a valid completed rating for your own report and document that convention. Before calculating, remove duplicates, invalid records, and technically incomplete responses that do not contain a usable answer to the question being scored.

A useful reporting note might say that the calculation uses valid answers to the satisfaction question, excludes skips and invalid records, and counts the top two choices as positive. Apply the same convention each time. If the rules change, mark the change rather than presenting the new result as directly comparable to the old one.

Read CSAT with response rate, timing, and sample size

CSAT describes the ratings you received; response rate describes participation among people eligible to answer. Calculate response rate using valid responses divided by eligible survey invitations or views, then multiply by 100. Consider who answered, when they were asked, and how many responses support a result before making decisions.

A survey response rate and a CSAT percentage have different denominators. Response rate is commonly calculated as:

Response rate (%) = (valid responses ÷ eligible invitations or views) × 100

CSAT instead uses positive ratings divided by valid completed ratings for the question. Keep these figures separate. A high CSAT percentage does not tell you what share of eligible customers responded, and response rate alone does not tell you whether respondents were satisfied.

A low response rate does not automatically prove the results are biased. It is a reason to look at who answered and who did not. For example, check whether feedback comes mostly from one channel, type of issue, or customer group. If the answers do not reflect the experiences you are trying to understand, a single aggregate score can mislead.

Timing affects what customers are able to recall. Ask about an interaction while it is still fresh enough for the respondent to assess. For a support team, that means keeping the question tied to the exchange being measured rather than asking much later about a vague overall experience. Avoid interpreting a score without noting when the survey was sent.

Small samples need cautious interpretation, especially when split into segments. A shift in a small group may reflect a few individual ratings rather than a broad change in service. Before acting on a segment, check its response count, compare similar periods and groups, and read the comments. Do not treat a precise-looking percentage as proof that the result is stable.

Calculation and survey-design mistakes that distort CSAT

CSAT becomes difficult to interpret when teams change scales, positive-response cutoffs, or denominator rules without saying so. Long surveys can also encourage people to repeat the same answer rather than consider each question carefully. Keep the method stable, ask focused questions, and remember that a company-wide score can hide differences between support channels.

One frequent problem is comparing percentages built from different rules. A result based on ratings of 4 and 5 is not directly equivalent to one based only on the top rating. Likewise, results from different scales or differently worded questions may not describe the same experience. Keep the question, response options, and scoring convention consistent when tracking a trend.

Another mistake is dropping neutral answers from the denominator to make the positive share look higher. If neutral answers are valid responses, excluding them changes the question the percentage answers. Decide how neutral responses are handled before reporting, and apply that rule consistently. Skips and invalid records can be excluded from the rating denominator, but should not be silently mixed with neutral ratings.

A questionnaire that asks too much can invite rushed or repetitive answers. Keep the survey focused on the experience you want to measure. If you need to understand several parts of a support interaction, consider whether each question has a clear purpose and whether its result will be reviewed or acted on.

An overall score can hide differences by channel, issue type, store, or operating period. Break results into segments when there are enough responses to make the comparison useful. A small segment can be informative as a signal to investigate, but avoid treating it as a settled conclusion.

Finally, CSAT does not tell the whole story of a customer relationship. A person may rate a single conversation positively while still having an unresolved broader issue, or rate a difficult interaction poorly despite a thoughtful response. Use the score as one input, then read comments and review the service context before deciding what to change.

Next steps: track trends, investigate comments, and make changes

Once the calculation is set, repeat it using the same question, scale, cutoff, and denominator rules. Compare like with like across periods, then read comments and review segments where the response count supports a useful comparison. Connect patterns to service behavior, address recurring issues, and measure again to see whether the experience changed.

Start by writing down the measurement definition: the question asked, who was invited, the response scale, which ratings count as positive, and how skips or invalid answers are handled. Keep that definition with the report. A stable method makes it easier to interpret movement without mistaking a change in the calculation for a change in customer experience.

Next, compare results over time using the same method. If the percentage moves, look at the number of valid answers as well as the percentage. Then read comments for explanations, while remembering that a few comments do not necessarily represent every respondent.

Use segments to locate a possible problem, not just to produce more numbers. A channel or issue-type split can point to a place worth reviewing when the sample is adequate. Compare feedback with operational details such as resolution behavior or communication, and look for a specific change the team can make.

For a support team, the practical follow-through is to review conversations connected to recurring feedback. Look for patterns such as unclear explanations, missed expectations, or a process that repeatedly creates confusion. In momo, teams can review conversations in the shared inbox and use feedback to identify issues worth addressing; it does not calculate CSAT or provide a survey feature.

After making a change, use the same calculation and question to check what happens next. Keep comments and service context in the review so the team can distinguish a meaningful pattern from an isolated rating. The point is not to chase a number at any cost; it is to understand what customers are reacting to and improve the experience.

Frequently asked questions

How to calculate CSAT score out of 5: is 4 the same as 80% CSAT?

No. A rating of 4 out of 5 is one respondent’s answer, and an average of 4 out of 5 summarizes a group’s ratings. Percentage CSAT is the share of valid ratings that meet the positive cutoff. If the cutoff is 4 or 5, calculate that share separately.

Do neutral responses count as satisfied in CSAT?

Usually, a neutral rating is not counted as positive when the positive cutoff is 4 or 5 on a five-point scale. It remains among valid answered ratings in the denominator. State your rule, since scoring conventions can differ, and use it consistently.

Should skipped survey questions be included in the CSAT denominator?

No, a skipped rating is not a valid answer to the satisfaction question, so exclude it from that question’s rating denominator. Track survey starts or invitations separately if you want to understand participation. Define “valid response” clearly in your reporting.

What is a good CSAT score?

There is no single threshold that makes a score good for every small support team. The result depends on the question, scoring method, respondents, and type of support experience. Establish a consistent baseline, compare like with like, and investigate comments and service patterns rather than judging the team by one percentage.

Can CSAT be calculated from a thumbs-up or star-rating survey?

Yes, if you define what counts as positive and calculate the share of valid responses meeting that rule. For a thumbs-up survey, that could be thumbs-up answers divided by valid answers. With stars, specify whether positive means the top rating or a wider set of ratings, and label the result accordingly.

Put the feedback into your support workflow

A consistent calculation helps you spot patterns; reviewing the conversations and comments helps explain them. If you are exploring AI support alongside a human inbox, try momo free.