DSAT, short for Dissatisfaction Score, is a customer service metric that measures the percentage of survey responses that rate an interaction negatively. It is the mirror image of CSAT and uses the same survey data: you count the responses that fall in the dissatisfied range and divide by the total number of responses. The formula is DSAT = (dissatisfied responses / total responses) x 100. A DSAT of 12% means 12% of the customers who answered your survey said the experience was poor. The term is used most often in contact centers and BPOs, where dissatisfaction is tracked as a target in its own right rather than read off the bottom of a CSAT report.
In short
- The DSAT formula is the number of dissatisfied responses divided by total responses, multiplied by 100.
- The full form of DSAT in customer service is Dissatisfaction Score, sometimes written as dissatisfaction rate or DSAT percentage.
- What counts as dissatisfied depends on your scale and your threshold, so two teams can report different DSAT numbers from identical survey data.
- DSAT and CSAT do not add up to 100%, because neutral responses sit between them and unanswered surveys sit outside both.
- DSAT is tracked separately from CSAT because a small, concentrated group of unhappy customers can be invisible in a healthy-looking average.
- A DSAT score tells you that a customer was unhappy. It does not tell you why, which is what makes the conversation behind the score the more useful record.
How to calculate DSAT
DSAT is a survey-based metric, calculated from the same post-interaction survey that produces your CSAT score. After a ticket, chat or call closes, the customer is asked to rate the experience. You count how many of those ratings fall in the dissatisfied range, divide by the total number of responses you received, and multiply by 100.
Note the denominator: it is total responses, not total surveys sent. Customers who never answered are not counted as dissatisfied, and they are not counted as satisfied either. They are simply not in the calculation at all, which is a limitation worth remembering when you read the number.
| Element | Detail |
|---|---|
| Formula | DSAT = (dissatisfied responses / total responses) x 100 |
| Dissatisfied responses | Ratings in the negative range, for example 1 and 2 on a 1 to 5 scale |
| Denominator | Total responses received, not total surveys sent |
| Scale | Usually 1 to 5, sometimes 1 to 3, or a simple thumbs up and thumbs down |
| Worked example | 18 dissatisfied out of 150 responses = 12% DSAT |
| Timing | Collected immediately after the interaction closes |
| Direction | Lower is better, unlike CSAT where higher is better |
What counts as a dissatisfied response
This is the part most definitions skip, and it is the part that changes your number. DSAT has no universal threshold. You choose which ratings count as dissatisfied, and that choice is a business decision, not a mathematical one.
On a 1 to 5 scale
The common convention is to count 1 and 2 as dissatisfied, 3 as neutral, and 4 and 5 as satisfied. This mirrors the standard CSAT convention of counting only the top two ratings as positive. Some teams count 3 as dissatisfied too, on the argument that a neutral rating after a support interaction is not a good outcome. That single decision can roughly double the reported DSAT on the same survey data.
On a 1 to 3 scale
A three-point scale, often labeled good, neutral and bad, leaves less room for interpretation. The bottom rating is dissatisfied. But it also compresses the signal: a customer who was mildly annoyed and a customer who was furious land in the same bucket, so the score tells you less about severity.
On a thumbs up and down survey
A binary survey removes the threshold question entirely, because there is no neutral option. DSAT becomes simply the share of thumbs down. The tradeoff is that you lose the ability to see the middle of your distribution shifting before it turns negative.
Why the threshold has to be written down
Because the threshold is a choice, a DSAT number is only comparable to another DSAT number when both were calculated the same way. If you benchmark against an external figure without knowing which ratings that team counted as dissatisfied, you are comparing two different metrics that share a name. Document your threshold, keep it stable, and treat your own trend over time as the real signal.
Why DSAT is not simply the inverse of CSAT
It is tempting to assume that if CSAT is 80%, then DSAT is 20%. It almost never is, for two reasons.
Neutral responses sit in between
On a five-point scale, a rating of 3 is usually counted as neither satisfied nor dissatisfied. Those responses are in the denominator of both calculations but the numerator of neither. So CSAT plus DSAT will typically fall short of 100%, and the gap is your neutral population. That gap is informative on its own: a large neutral block means a lot of customers came away feeling nothing much, which is a different problem from a small block of angry ones.
Non-responses sit outside both
Response rates on post-interaction surveys are typically low, and the customers who answer are not a random sample. People with a strong reaction in either direction are more likely to reply than people who were mildly inconvenienced. That means both CSAT and DSAT are computed on a self-selected slice of your customers, and the dissatisfaction that never gets a survey response is invisible in both numbers.
The practical consequence: read DSAT as the share of respondents who told you they were unhappy, not as the share of customers who were unhappy. Those are different populations. The full breakdown of how the two metrics diverge on the same data is in CSAT vs DSAT.
What a DSAT trend tells you that CSAT hides
CSAT is an average, and averages absorb small groups. A support operation can hold a steady 82% CSAT for two quarters while the share of genuinely angry customers quietly doubles, because the growth in 1 and 2 ratings is offset by movement between 3s, 4s and 5s. Tracked on its own, DSAT surfaces that movement immediately.
The reason to watch it separately is that dissatisfied customers behave differently from merely unsatisfied ones. They escalate, they churn, they write reviews, and they generate repeat contacts. The cost of a support operation is concentrated in that tail, so the tail deserves its own metric.
Read DSAT as a distribution, not a single number
The most useful view is not this month’s DSAT percentage but where the dissatisfied responses cluster: by queue, by channel, by issue type, by time of day, by tenure of the agent handling it. Dissatisfaction is rarely spread evenly. When you segment it, a flat-looking 9% often turns out to be 4% almost everywhere and 30% in one workflow that nobody has looked at.
Pair it with the metrics that explain it
DSAT tends to move with the operational metrics around it. A rise often traces back to first response time, first contact resolution, or the number of transfers before a resolution. Reading DSAT alongside those, rather than alone, is what turns it from a scoreboard into a diagnosis. The wider set of metrics that belong on the same dashboard is covered in our guide to customer service metrics and KPIs.
DSAT in the call center and BPO context
DSAT appears far more often in contact center and BPO vocabulary than in general CX writing, and there is a structural reason for that. In an outsourced or high-volume environment, quality is contractual. A BPO is frequently measured against a service level agreement that sets a maximum tolerable dissatisfaction rate, not just a minimum satisfaction score, because the client cares about the floor of the experience as much as the average.
That framing changes how the metric is used. In-house CX teams often treat DSAT as a diagnostic they look at when something feels wrong. Contact centers treat it as a managed target with a threshold, an owner, and a review cadence. Each dissatisfied response is typically expected to be reviewed individually, categorized, and either accepted as outside the team’s control or routed into coaching.
It also explains the vocabulary you will encounter around it: DSAT tickets, DSAT reviews, DSAT drivers, DSAT recovery. These all describe the same underlying workflow of taking the negative responses seriously one at a time rather than watching a percentage move.
From a DSAT score to the reason behind it
The limitation of DSAT is the same as the limitation of every survey metric: it records an outcome without recording a cause. You know a customer rated the interaction badly. You do not know whether the agent was unclear, the policy was unpopular, the answer was wrong, the wait was long, or the customer was already frustrated before the conversation started. Those require entirely different responses, and the score cannot tell them apart.
The cause is in the conversation itself. Every dissatisfied rating has a transcript attached to it, and that transcript contains the actual sequence of events. Reading those conversations, rather than reading the score, is how a DSAT number becomes an action. Doing it at scale, and categorizing what you find rather than collecting anecdotes, is the subject of how to reduce DSAT.
This is also where quality assurance and survey data meet. QA scores what happened inside the conversation against your own standards for accuracy, process and tone. Survey metrics record what the customer felt about it afterward. Kaizo measures that internal quality and traces every score back to the evidence in the transcript, so a dissatisfied rating can be connected to the specific handling behavior behind it rather than debated. Because it runs natively in Zendesk and Salesforce, the conversation and the score sit in the same place.
Frequently asked questions
What is the DSAT formula?
DSAT equals the number of dissatisfied responses divided by the total number of responses, multiplied by 100. If 18 out of 150 survey responses fall in your dissatisfied range, your DSAT is 12%. The denominator is responses received, not surveys sent, so customers who never answered are excluded from the calculation entirely.
What is the full form of DSAT in customer service?
DSAT stands for Dissatisfaction Score, sometimes written as dissatisfaction rate. It is the counterpart to CSAT, which stands for Customer Satisfaction Score. Both are calculated from the same post-interaction survey. Note that outside customer service, the same four letters are used for an entirely unrelated standardized test, so context matters.
What counts as a dissatisfied response?
You choose the threshold. On a 1 to 5 scale the common convention is to count ratings of 1 and 2 as dissatisfied, with 3 treated as neutral. Some teams include 3 on the argument that a neutral outcome after a support contact is still a failure, which can roughly double the reported number on identical data. Whichever you pick, write it down and keep it stable, because a DSAT figure is only comparable to another one calculated the same way.
Is DSAT just the opposite of CSAT?
No. CSAT and DSAT do not add up to 100%. Neutral responses are counted in the denominator of both but the numerator of neither, so they create a gap between the two figures. On top of that, both metrics are calculated only on customers who responded, and response rates are typically low, so dissatisfaction that never produced a survey response is missing from both.
What is a good DSAT score?
There is no credible universal benchmark, because the number depends entirely on your scale, your chosen threshold, your survey timing and your response rate. A team counting neutral ratings as dissatisfied will report a much higher DSAT than a team that does not, on exactly the same customers. The useful comparison is your own trend over time and the variation between your own queues, channels and issue types.
Why do contact centers track DSAT separately from CSAT?
Because an average absorbs a small group. CSAT can hold steady while the share of genuinely angry customers grows, since movement between the middle ratings offsets it. Dissatisfied customers also drive disproportionate cost through escalations, repeat contacts and churn, and in outsourced environments a maximum dissatisfaction rate is often written into the service level agreement as a target in its own right.
Related terms
See the conversations behind your DSAT score
A dissatisfaction score tells you a customer was unhappy. The transcript tells you why. Bring a week of your real conversations and we will show you the quality signal behind your negative ratings, with every score traced back to the evidence in the ticket.
