What is net promoter score?
NPS summarizes responses to a likelihood-to-recommend question on a scale from zero to ten. Bain’s Net Promoter System groups ratings of nine or ten as promoters, seven or eight as passives, and zero through six as detractors. Subtracting the detractor share from the promoter share gives the score. Passives remain in the response denominator.
For a SaaS business, the result can organize a conversation about the customer relationship. It does not explain why someone chose a rating, whether they actually referred a colleague, or whether they will renew. Read the score alongside the customer’s explanation and observed product behavior rather than converting it directly into a revenue forecast.
NPS calculator
Enter counts from one survey wave using the same eligibility and response rules. A respondent belongs in exactly one group. Exclude missing ratings from the score, and record the number of skipped questions separately. An empty survey has no NPS; it is different from a completed survey whose promoters and detractors balance to zero.
The fictional defaults below use 20 responses. The calculator handles arithmetic only: it does not collect responses, validate your sample, or estimate the views of people who did not answer. Keep identifiable survey records in your own appropriate research system.
Enter response counts from one defined survey wave. The default is fictional. Calculations stay in your browser; no responses are submitted.
- Net promoter score
- 45.00
- Total responses
- 20
- Promoters
- 60.00%
- Detractors
- 15.00%
JavaScript is required to update the calculator. The formula and worked example below remain readable without it.
NPS formula and a worked SaaS example
NPS = (promoter responses − detractor responses) ÷ all valid responses × 100. The equivalent calculation is percentage promoters minus percentage detractors. The scale runs from −100 when every respondent is a detractor to +100 when every respondent is a promoter. A score of 45 is written as NPS 45, not 45%.
Consider an original fictional wave with 12 promoters, five passives, and three detractors. There are 20 valid responses. Promoters represent 60%, passives 25%, and detractors 15%. NPS is 60 − 15 = 45. Removing passives from the denominator would produce 60 instead, answering a different calculation.
Suppose 80 eligible people were invited and those 20 answered. The response rate is 25%, and the views of the other 60 people remain unknown. If one passive respondent had instead become a promoter, the score would rise by five points. That sensitivity matters when interpreting small survey waves. These counts are invented teaching data, not ShowMeYourSaaS results.
| Group | Rating | Count | Share of valid responses |
|---|---|---|---|
| Promoters | 9–10 | 12 | 60% |
| Passives | 7–8 | 5 | 25% |
| Detractors | 0–6 | 3 | 15% |
| All valid responses | 0–10 | 20 | 100% |
Choose relationship or experience feedback deliberately
Bain distinguishes feedback about the overall relationship from feedback triggered by a particular experience, as well as competitive benchmark research. A survey after a support conversation is framed differently from a periodic review of the whole product relationship. Keep those purposes visible when comparing scores.
For a fictional monthly reporting SaaS, an overall relationship wave might include customers who have had an opportunity to complete a reporting cycle. A separate experience survey could investigate a completed import. Do not combine the two waves just to increase the sample count. Record the product version, eligibility window, invitation channel, and any changes in wording or timing.
Set a contact frequency appropriate to the relationship. If the same person receives overlapping product, support, and renewal surveys, both the response burden and the context change. Keep an invitation log and coordinate follow-up before sending another request.
Who should answer a B2B SaaS survey?
An organization can contain a buyer, an administrator, and several end users with different experiences. Bain’s B2B guidance highlights that complexity. Decide which role can answer the question you are investigating, then state whether your score represents people or accounts.
If one company contributes ten responses and another contributes one, a respondent-weighted score gives the first company more influence. That may fit an end-user study, but it is different from giving each customer account equal weight. Do not silently average a mixture of the two approaches. Show role and account coverage alongside totals.
Start with data you actually need. A product role and a recent completed job may explain a response better than collecting additional personal characteristics. Give customers a clear explanation of how feedback will be used and make a follow-up conversation optional.
What should the survey ask after the score?
Keep the standard score question and endpoints consistent when tracking that method. The download includes the attributed question and our original follow-up prompts. A useful first follow-up asks for the experience behind the rating, without steering the respondent toward a feature or a positive explanation.
For example, ask what happened the last time they tried the core task, what helped, and what remained difficult. Ask permission before moving from an anonymous response to an identified support conversation. If someone reports an unresolved failure, the immediate next step may be helping them complete the task rather than asking for a public review.
- What experience most influenced the rating you chose?
- What were you trying to accomplish on your most recent visit?
- What, if anything, made that task harder than it needed to be?
- What should we preserve because it already works well for you?
- Optional: may we contact you to understand this response?
What is a good NPS, and what can the score miss?
Use comparisons only when the audience, survey type, collection method, and timing are comparable. A vendor’s headline benchmark is not automatically an appropriate target for a small, newly launched SaaS. Start by making your own waves reproducible, then investigate changes with the underlying comments and counts.
The same score can describe different distributions. Ten promoters and ten detractors produce zero; twenty passives also produce zero. The first group is polarized, while the second contains no strongly favorable ratings under the scoring rule. Keep the three groups visible instead of retaining only the final score.
NPS also answers a different question from product-market fit, churn, or successful task completion. A person can recommend a tool while having no current reason to use it again. Check the customer’s natural workflow cadence and actual renewal behavior before interpreting a score as retained demand.
Download the NPS survey and review worksheet
The editable worksheet includes the attributed core item, original follow-ups, eligibility rules, the fictional count example, and a follow-up decision log. Method sources were checked on October 26, 2026. NPS terminology is attributed to Bain’s Net Promoter System; this calculator is an independent educational implementation.