Most B2B teams send an NPS survey, get a number, and then argue about whether it is "good". The score is the least useful part. This guide shows how to calculate NPS correctly, choose who to survey and when, and build the follow-up routine that turns answers into saved renewals.

What NPS is and how to calculate it

Net Promoter Score asks one question: "How likely are you to recommend us to a colleague or friend?" Respondents answer on a 0 to 10 scale and fall into three groups:

  • Promoters: 9 or 10
  • Passives: 7 or 8
  • Detractors: 0 to 6

The score is the percentage of promoters minus the percentage of detractors, which gives a number between -100 and +100 (Survicate's explanation). Passives are left out of the subtraction but still count in the total, so they pull the score toward zero.

NPS = % promoters - % detractors

Example (illustrative): 50 responses
  22 promoters (44%), 20 passives (40%), 8 detractors (16%)
  NPS = 44 - 16 = +28

Report the score as a whole number, not a percentage, and always report the number of responses next to it. A +28 from 8 responses means very little.

What is a good NPS for B2B SaaS?

Sources disagree, which is the first sign you should not obsess over the benchmark. Enterpret's 2026 guide puts the B2B SaaS median at roughly 36, with published figures ranging from about 30 to 41, and treats 40 or higher as good and 50 or higher as excellent. It also reports that scores rise with company size: about 25 for companies under $1M ARR, 32 for $1M to $10M, and 35 above $10M. CustomerGauge's ten-year review reports a cross-sector B2B benchmark of 48, a reminder that the number depends on the industry and the panel.

Two practical conclusions:

  1. Compare yourself with your own previous surveys first. Enterpret's advice is the same: consistent methodology across waves is more reliable than an industry median.
  2. Who you ask changes the result. Executive buyers tend to score higher than daily end users, so mixing them without noting it can hide a problem.

Step 1: Decide who to survey

In B2B, one account has several people with different views. Decide this before you send anything.

  • Relationship NPS goes out to a defined set of contacts once or twice a year (for example, every six months). It measures overall sentiment.
  • Transactional NPS or CSAT goes out after a specific event, such as onboarding completion or a closed support ticket. It measures that experience.

Pick the roles that matter: the economic buyer, the champion and the main day-to-day users. Survey the same roles each wave. Skip contacts who are mid-escalation or in the first 30 days of onboarding, because their answer reflects a single moment.

Step 2: Send it in a way people answer

Keep it to the 0 to 10 question plus one open text question. CustomerGauge reports a median response rate of 19% on its platform (15% in its self-reported survey), and that 40% of companies have a contact response rate of 10% or less. Its data also suggests that programs with a formal response-rate target do better: 49% of those programs beat 20%, versus 30% without one.

A few things that help:

  • Send from the CSM's name, not a no-reply address.
  • Send Tuesday to Thursday, mid-morning in the recipient's time zone.
  • Tell people the survey takes one minute and who reads the answers.
  • Follow up with a short nudge a week later, and then stop.

Use a different open question depending on the score, as many programs do: "What do we do well?" for promoters, "What would make us a 9 or 10?" for passives, and "What went wrong?" for detractors.

Step 3: Close the loop

This is where NPS earns its place. Bain, the origin of the method, says that organizations that use NPS to create closed-loop customer feedback learn faster how to create more promoters. In its example, employees call detractors who agreed to be contacted, plus a sample of promoters, and then fix the problem or hand it to someone who can.

Speed matters. CustomerGauge reports that 62% of programs now set close-the-loop targets (up from 30% in 2021) and that a 48-hour response is becoming the norm, yet only 38% actually close the loop within 48 hours.

Use a simple routing table:

Score      Owner          Action                           Deadline
0-6        CSM            Call or personal email; log      2 business days
           (+ manager     root cause; create a task
           if ARR is      for any fix
           high)
7-8        CSM            Short thank-you; ask what        1 week
                          would make it a 9 or 10
9-10       CSM            Thank; ask for a reference,      1 week
                          review or introduction
Score drop of 3+ points since last wave:
           CSM + manager  Treat as a risk signal,          2 business days
                          review at next churn review

A detractor call script

1. Thank them for the honest answer. Do not defend.
2. "You gave us a 4 and wrote [quote]. Can you tell me more?"
3. Ask: "What would need to change for this to work for you?"
4. Repeat back what you heard. Agree one next step and a date.
5. Log it on the account, create the task, and tell them when it is done.

The last step is the one teams skip. Tell the customer what changed. That follow-up is often what converts a detractor into a renewal.

Step 4: Look beyond the average

A blended score can hide a declining segment. Enterpret recommends segmenting by ARR, tenure and respondent type, and grouping open-text comments into recurring themes weighted by the accounts and revenue involved. A theme named by three small accounts matters less than one named by two accounts that represent a fifth of your ARR.

A quick monthly view:

  • NPS and response count by segment (size, tier, tenure)
  • Top three detractor themes, with ARR attached
  • Detractor loops closed within target, and how many are still open
  • Accounts whose score dropped since the last wave
  • Promoters with no reference or case study request yet

Enterpret also notes that a customer whose score falls, for example from 9 to 6, is a churn signal whatever the industry median is. Feed NPS into your health score as one signal among several instead of treating it as a verdict; our customer health score formula guide shows how to weight it. CustomerGauge adds a caution on revenue: it found that only 4.5% of programs have quantified the financial impact of customer experience, so linking scores to renewal outcomes in your own data is a real advantage.

Common mistakes

  • Surveying only the champion. The person who signs the renewal may feel differently.
  • Changing the question or panel between waves. You lose the trend.
  • Reporting the score without the response count.
  • Treating passives as a non-event. They are the easiest group to move upward and the easiest to lose.
  • No owner for detractors. If the answer lands in a shared inbox, nobody calls.

Try it in XPI CRM

XPI CRM includes NPS and CSAT surveys, and NPS is one of the seven factors in the 0 to 100 health score, with alerts for detractors so a low score does not sit unread. You can attach a follow-up task to the account, or start a playbook automatically, and see the survey history next to renewals and support tickets. See the help guides, or compare it with tracking this in spreadsheets. Start free: every feature included, no credit card.