NPS and CSAT answer different questions. Net Promoter Score (NPS) asks how likely someone is to recommend you, so it tracks the overall relationship over time. Customer Satisfaction Score (CSAT) asks how happy someone is with one specific interaction, so it tells you where a particular touchpoint is working or failing. If you have to pick one, pick by the decision you need to make. If you can run both, run both, and add a third measure, Customer Effort Score (CES), for the moments where friction drives churn.
This guide is written for B2B SaaS product and growth teams. It covers the formulas, the benchmarks, the traps, and a practical setup for a free-trial product, including how product qualified leads (PQLs) fit in, since many readers land here from that question too.
TL;DR
- NPS is a relationship metric: one 0-10 question about recommending you, scored from -100 to +100. Use it quarterly or twice a year to see the health of the account.
- CSAT is a transactional metric: a satisfaction rating after a specific moment, reported as a percentage. Use it right after onboarding steps, support tickets or feature use.
- They are not rivals. They survey different people at different times, so comparing one number to the other is a mistake.
- CES often catches churn risk earlier than either. In a free trial, effort during setup is usually the signal worth watching.
- A PQL is a user who has shown buying intent through product behavior. Survey scores can sit next to usage data, but they do not replace it.
Table of Contents
- What NPS measures and how it is calculated
- What CSAT measures and how it is calculated
- NPS vs CSAT at a glance
- Why you cannot compare the two scores directly
- What counts as a good score
- Where CES fits: NPS vs CSAT vs CES
- How to choose: a decision guide
- What this looks like in a B2B free trial
- PQLs: what they are and how they relate to these metrics
- Our perspective: measure friction before you measure feelings
- Getting help turning scores into fixes
- Common mistakes to avoid
- FAQ
- Sources
What NPS measures and how it is calculated
NPS was developed in 2003 by Fred Reichheld of Bain & Company to gauge loyalty. The survey asks one question: how likely are you to recommend us to a friend or colleague? Answers run from 0 to 10.
- Promoters: 9 to 10
- Passives: 7 to 8 (left out of the math)
- Detractors: 0 to 6
The formula is the percentage of promoters minus the percentage of detractors. The result is a whole number between -100 and +100, not a percentage.
A quick example. You survey 200 customers. 90 answer 9 or 10, 70 answer 7 or 8, and 40 answer 0 to 6. Promoters are 45%, detractors are 20%, so your NPS is +25.
Two details matter in practice. First, the same source notes that the 0-10 scale was refined from an original 1-10 scale so that zero marks a distinct bottom. Second, passives are not a safe middle. Practitioners treat 7s and 8s as at-risk, because they are unlikely to drive the organic growth that promoters do.
What CSAT measures and how it is calculated
CSAT measures immediate satisfaction with a specific product, service or interaction. The usual question is some version of "How satisfied were you with X?" on a 1-5 scale. The calculation divides positive responses by total responses and multiplies by 100, giving a percentage. Most teams count 4s and 5s as positive.
Example: 120 people rate their onboarding call. 84 answer 4 or 5. CSAT is 84 / 120 x 100 = 70%.
Because CSAT attaches to a moment, it is the metric for operational feedback: closing the loop on a support issue, checking a new onboarding checklist, or testing whether a redesigned screen helped. It moves quickly, and it can swing on a single bad day.
NPS vs CSAT at a glance
| NPS | CSAT | |
|---|---|---|
| Question | How likely are you to recommend us? | How satisfied were you with this? |
| Scale | 0-10 | Usually 1-5 |
| Result | -100 to +100 | 0-100% |
| Focus | Long-term loyalty and sentiment | Short-term satisfaction with one touchpoint |
| Who answers | Broad base, including quiet customers | People who just had the interaction |
| Timing | Periodic (quarterly, semiannual) | Event-triggered |
| Best for | Strategic health check, renewal risk, trends | Diagnosing specific steps, support quality |
| Weakness | Does not say why | Can miss silent, disengaged customers |
Why you cannot compare the two scores directly
The most common mistake is treating CSAT and NPS as interchangeable or trying to convert one into the other. As Sogolytics explains, they measure different things, on different scales, from different populations, at different points in the relationship. Transactional CSAT only hears from engaged users after a touchpoint. Relationship NPS reaches a broader and often quieter base.
That is how a company ends up with strong CSAT and falling NPS. Support is excellent, so every ticket ends happily, while the product itself slowly frustrates people who never open a ticket. Experts call this the satisfaction trap: assuming high CSAT guarantees retention, when it may mask churn among silent customers.
The reverse happens too. A product people love can have a clumsy support flow and mediocre CSAT on tickets, while NPS stays healthy.
What counts as a good score
Benchmarks vary by industry, company size and survey method, so treat them as a rough range, not a grade.
NPS. Reichheld's framing, as summarized here, is that top performers typically land between +50 and +80, while the average company scores below +10. For SaaS specifically, SurveySparrow's 2026 roundup reports a median of about 30, with companies under $1M ARR averaging 25 and those above $10M averaging 35. Those figures come from one vendor's data set, so compare yourself to your own history first.
CSAT. There is no universal threshold. The 85% reported in a 2025 study of a small business's CRM rollout shows what a post-transaction score can look like, but it is a single case, not a SaaS norm. Set your own baseline per touchpoint and watch the direction.
If you saw the question of whether a 7 is a good NPS score: a 7 is a passive. It does not count toward promoters and does not count against you, but it is not a vote of confidence either.
Where CES fits: NPS vs CSAT vs CES
Customer Effort Score asks how easy it was to get something done. CEB Global Research, now part of Gartner, developed it in 2010, and the same source reports a Gartner finding that 96% of customers with a high-effort experience become more disloyal, against 9% of those with a low-effort one. It also cites CES as roughly 40% better than CSAT at predicting future purchase behavior. Treat those as Gartner-attributed figures from a customer service context, not guarantees for your product.
Gartner's own guidance argues that effective programs move away from delighting customers and toward an effortless experience: fewer repeat inquiries, more self-service, simpler workflows.
CES has limits. One vendor guide notes it only reflects the feeling right after a specific interaction, like CSAT.
| Question you are asking | Metric |
|---|---|
| Do customers feel loyal to us overall? | NPS |
| Was that specific step or ticket handled well? | CSAT |
| Was that task harder than it should have been? | CES |
For NPS vs CES specifically: NPS tells you how the relationship looks, CES tells you which part of the experience is costing you. They pair well. NPS is the headline, CES is the diagnosis.
How to choose: a decision guide
Start from the decision, not the metric.
- You need a board-level health number or renewal risk signal. Use NPS, sent to a broad sample on a fixed schedule.
- You want to know whether a new onboarding step or support process works. Use CSAT triggered right after it.
- You suspect users are struggling but not complaining. Use CES on the specific task: connecting an integration, importing data, inviting a teammate.
- You need to know why scores move. Add one open-text follow-up to every survey: "What's the main reason for your score?" The number says how much, the comment says why.
A few survey habits keep the data honest:
- Keep each survey to one rating question plus one open comment.
- Do not survey the same person with three instruments in a week.
- Segment results by plan, role and account size. A blended NPS hides a lot.
- Report response rate beside every score. A +40 from 12 responses is not a +40.
- Close the loop. Reach out to detractors and low CSAT respondents quickly.
What this looks like in a B2B free trial
Free trials are where the NPS vs CSAT question gets tricky, because most trial users never become customers and many never reach a moment worth rating.
A workable setup:
- CES at the setup milestone. After a user connects their data or completes the first key action, ask how easy it was. Setup that comes before any value is one of the most common places trials leak.
- CSAT after the first value moment. When a user sees their first result, ask how satisfied they are with it.
- NPS for active paying accounts, not for trial users who barely touched the product. Their answer would say little about the relationship.
- Behavior data beside all of it. Surveys only reach people who respond. Usage shows what the silent majority did.
That last point leads to PQLs.
PQLs: what they are and how they relate to these metrics
A product qualified lead is, in one practical definition, a user or account that has experienced meaningful value in your product and shown buying intent through usage, as opposed to someone who just signed up or downloaded a whitepaper. PQLs typically emerge in free trials, freemium models and sandboxes.
PQL vs MQL. A marketing qualified lead is scored on marketing engagement: content downloads, webinar attendance, email clicks. A PQL is scored on what the person did inside the product. MQLs signal interest. PQLs signal that the product has already delivered something.
PQL scoring. There is no standard formula. You build it from your own data: find the actions that paying customers took in their trial, such as inviting teammates, hitting a usage threshold, using a key feature several times, or visiting the pricing page after activation. Weight those actions, set a threshold, then test whether high scorers convert at a higher rate than the rest.
Survey scores can feed that model as a secondary input. A high CES at setup or a strong CSAT after the first result can confirm a good experience. But a user who gives a lukewarm NPS and still invites five colleagues is more of a PQL than a quiet promoter who never logs in. Behavior leads, sentiment supports.
Our perspective: measure friction before you measure feelings
We come at this from the product side, not the survey side. When we work with a SaaS team, we sign up for the trial as a real user, go through it step by step, and note where we stall: setup that asks for effort before delivering value, prompts that arrive at the wrong moment, pricing that is unclear. Almost every time, the scores a team already tracks point at the symptom, not the cause.
Here is what we have learned to say to product leaders.
A single number will not save you. NPS is a good thermometer and a bad diagnostic. CSAT is a good diagnostic for the moment it measures and blind to everything else. Neither tells you why a trial user left on day two without a word.
Effort is the metric most teams underuse. The research on effortless experience comes from customer service, but the logic carries over to product flows. Every extra field, every unexplained integration step, every empty dashboard is effort. We would rather you fix the first-run experience than lift your NPS by two points.
Put the survey where the decision is. A satisfaction question inside the screen that is losing users tells you more than a quarterly email to everyone.
Pair every score with watching real behavior. Session recordings, funnel data and a few user conversations tell you what the number cannot. Our free-trial teardowns work exactly this way: observation first, then research to prioritize what to fix.
If you want a sense of how we apply that kind of product-first thinking, look at our work with Supliful, where product setup changes supported $1.1M in GMV growth.
— Edijs
Getting help turning scores into fixes
If your trial-to-paid conversion is flat and your survey data does not explain why, the gap is usually between the metrics and the actual user experience. Our AI + UX Discovery engagement is built for that: we go through your trial as a user, map the drop-off points, and show you what to fix first and why. You can see everything we offer on the services page.
Common mistakes to avoid
- Using NPS on trial users who have not seen value. The score reflects confusion, not loyalty.
- Reporting CSAT without the denominator. Percentages from tiny samples mislead.
- Chasing the score. Incentivizing teams on a number invites gaming, such as surveying only happy customers.
- Ignoring passives. 7s and 8s are where the next churn or the next promoter comes from.
- Surveying only people who contact support. You will measure your support team, not your customer base.
- Skipping the open-text question. Without reasons, you have a scoreboard and no plan.
FAQ
Is NPS or CSAT better for SaaS?
Neither wins outright. NPS is better for tracking loyalty and renewal risk across your customer base, while CSAT is better for judging a specific step like onboarding or a support interaction. Most SaaS teams use NPS on a schedule and CSAT after key events.
How do I calculate NPS and CSAT?
For NPS, subtract the percentage of detractors (0-6) from the percentage of promoters (9-10). The result runs from -100 to +100. For CSAT, divide the number of positive responses, usually 4s and 5s, by total responses and multiply by 100.
What is a good NPS and CSAT score?
One SaaS benchmark report puts the median SaaS NPS near 30, and scores above 50 are generally called world-class. CSAT has no universal standard, so set a baseline for each touchpoint and track the trend. Your own history is a better yardstick than another company's average.
Can I convert NPS to CSAT?
No. They use different scales and survey different groups of people at different times, so there is no valid conversion. Run both and read them side by side.
What does PQL mean?
PQL stands for product qualified lead: a user or account whose in-product behavior shows they got real value and may be ready to buy. It differs from an MQL, which is based on marketing engagement. You define PQL scoring from the actions your paying customers took before they converted.
Sources
- NPS Promoters, Passives, Detractors: Score Ranges Explained
- CSAT vs NPS: Difference Between CSAT and NPS, Sogolytics
- Implementasi Customer Relationship Management dengan Indikator Customer Satisfaction Score
- How to Improve Customer Loyalty With Customer Effort Score
- Create an Effortless Customer Service Experience, Gartner
- What Is a Good Customer Effort Score?
- Improves Customer Effort Score
- SaaS NPS Benchmarks 2026
- What Is a Product Qualified Lead? PQL Guide for B2B