WiseData NPS works with three metrics: NPS, CSAT and CES. Which one a survey produces is decided by the type chosen in the first step: the six NPS types (Relational, Transactional, Product, eNPS, Brand and Competition) produce NPS, the CSAT type produces CSAT and the CES type produces CES. Each metric has its own scale and math, and the system never adds responses from different scales into a single number — a 5 is the best possible CSAT rating and would be a detractor in NPS.
Throughout this article we'll use Aurora Store, which measures NPS after each purchase and CSAT after each support interaction.
NPS: recommendation, from 0 to 10
The classic question is "how likely are you to recommend our company to a friend or colleague?", answered from 0 (Not at all likely) to 10 (Extremely likely). The rating puts each person in one of three groups:
- Promoters (9 and 10) — enthusiastic customers who tend to refer you and come back.
- Passives (7 and 8) — satisfied, but not enthusiastic.
- Detractors (0 to 6) — unhappy.
The math is NPS = % of promoters − % of detractors. Passives count toward the total but neither add nor subtract — which is why a survey with only passives has an NPS of 0. The result ranges from −100 to +100. If Aurora Store got 200 responses, with 120 promoters (60%), 60 passives (30%) and 20 detractors (10%), the NPS is 60 − 10 = 50.
CSAT: satisfaction, from 1 to 5
Measures satisfaction with a specific support interaction or delivery, from 1 (Very dissatisfied) to 5 (Very satisfied). The ranges are Satisfied (4 and 5), Neutral (3) and Dissatisfied (1 and 2).
The math is CSAT = % of responses rated 4 or 5, from 0% to 100%. It is not the average rating: if half the customers give 5 and half give 1, CSAT is 50%, while the average (3) would suggest a lukewarm service that doesn't exist. If 80 of the 100 customers Aurora Store helped gave 4 or 5, CSAT is 80%.
CES: effort, from 1 to 7
Measures how easy it was to solve the problem. The person says whether they agree the company made it easy to solve, from 1 (Strongly disagree) to 7 (Strongly agree) — so a high rating is good, just like the other metrics. The ranges are Low effort (6 and 7), Moderate effort (3 to 5) and High effort (1 and 2).
The math is the average of all ratings, with one decimal place (for example, 5.6). The closer to 7, the easier it was for the customer.
The rating of each result
Next to the number, the screen shows a rating. The cut-offs differ by metric, because what's great for CSAT isn't great for NPS:
- Excellent — NPS from 75; CSAT from 80%; CES from 6.0.
- Good — NPS 50 to 74; CSAT 76% to 79%; CES 5.5 to 5.9.
- Fair — NPS 25 to 49; CSAT 65% to 75%; CES 4.5 to 5.4.
- Poor — NPS 0 to 24; CSAT 50% to 64%; CES 3.5 to 4.4.
- Critical — NPS below 0; CSAT below 50%; CES below 3.5.
The same cut-offs apply on screen and in exported reports. In the example, Aurora Store's NPS of 50 is Good and its 80% CSAT is Excellent.
How to read it without fooling yourself
- Watch the trend. An NPS that goes from 30 to 45 in three months says more about your operation than comparing the number with another company's.
- Be wary of small samples. With 10 responses, each person weighs 10 points; one more response changes the rating.
- Compare like with like. NPS, CSAT and CES aren't comparable with each other — each has its own scale.
- Group percentages may add up to 99% or 101%. Each one is rounded separately for display; the score is calculated on the exact counts, so it's correct.
- The number tells you how much; the comment tells you why. Follow-up questions, set up by score range, are what explain a drop.
Where to go next
- Survey types — which type to use for each question, and which metric it produces.
- Results and reports — where to follow the score, the distribution and the evolution.
- Alerts — getting warned when NPS drops below a value.
In short: the survey type defines the metric — NPS (% of promoters minus % of detractors, from −100 to +100), CSAT (% of 4 and 5 ratings) or CES (average from 1 to 7) —, each with its own ranges and Excellent to Critical rating; read the trend, be wary of small samples and don't compare different metrics.