New webinar on 3rd March 2026 | 9:00 am PT
In this webinar, discover how Sopact Sense revolutionizes data collection and analysis.
NPS = % Promoters − % Detractors. Complete formula, worked example, 2026 industry benchmarks, and the system that makes the number actually matter.
Your NPS came back as +42. Leadership called it "solid." Three weeks later, your highest-value customer churned — an account that had scored a 6 on that same survey. Nobody followed up because the survey was anonymous. Nobody flagged the account because the score was reported as one number. The formula worked. The program failed.
Last updated: April 2026
The NPS calculation formula is NPS = % Promoters minus % Detractors, where Promoters score 9–10 and Detractors score 0–6 on a 0–10 recommendation scale. Passives (7–8) are counted in the total but excluded from the subtraction. That math takes thirty seconds. The system that connects each score to a specific customer, segment, and touchpoint — and triggers action before churn — is what actually moves the score. This guide covers both: the formula and benchmarks first, then the program design that makes the number matter.
The NPS calculation formula is NPS = % Promoters − % Detractors. Promoters are respondents who score 9 or 10 on the "How likely are you to recommend us?" question. Detractors score 0–6. Passives (7–8) count toward the response base but are excluded from the subtraction. The resulting score ranges from −100 (all detractors) to +100 (all promoters) and is reported as a whole number, not a percentage.
The formula does not change across industries, response volumes, or NPS types (customer NPS, employee NPS, beneficiary NPS). What changes is the architecture around it — the touchpoints where you collect, the segments you can disaggregate by, and the speed at which signal reaches the person who can act on it.
Calculate NPS in four steps: (1) classify every respondent as Promoter (9–10), Passive (7–8), or Detractor (0–6); (2) divide Promoters by total respondents to get Promoter %; (3) divide Detractors by total respondents to get Detractor %; (4) subtract Detractor % from Promoter %. The result is your NPS, expressed as a whole number between −100 and +100.
The Passive count affects your denominator but never your numerator. This is the most common arithmetic mistake in NPS calculation — teams either forget to include Passives in the total (which inflates their score) or they subtract Passives from Promoters (which is not the formula). Only Promoters and Detractors enter the subtraction.
Out of 200 survey respondents: 120 scored 9 or 10 (Promoters), 40 scored 7 or 8 (Passives), and 40 scored 0–6 (Detractors). Promoter % = 120 ÷ 200 = 60%. Detractor % = 40 ÷ 200 = 20%. NPS = 60 − 20 = +40. The Passives are visible in the total but never enter the subtraction.
Two edge cases matter. First, the 1–5 scale that survey tools sometimes default to is not NPS — it produces a different distribution and the formula does not apply without a validated transformation. Second, statistical significance collapses below 50 responses: your score can shift 10+ points from random noise, which is why quarterly programs with small cohorts frequently report "movement" that reflects sample variance, not sentiment change.
Any NPS above 0 means more respondents would recommend you than warn others away. Scores above +50 are excellent. Scores above +70 are world-class and rare — achieved by companies like Apple and USAA in specific segments. The benchmark that matters most is your own trend over time, disaggregated by segment, not a competitor's published aggregate.
Industry context shifts the baseline substantially. A +20 in cable or telecommunications is strong; the same +20 in e-commerce or SaaS signals a problem. Published "industry averages" are worth less than they appear — they average customer NPS, employee NPS, transactional NPS, and relational NPS into a single headline number that nobody's program actually measures. Your trend against your own prior quarter, broken out by segment, is the only benchmark that drives decisions.
The structural problem with traditional NPS reporting is that the formula collapses everything into one integer. You are averaging together the loyalty of customers who just onboarded, customers who have used your product for three years, enterprise accounts worth $200K, and trial users on day seven. Then leadership asks the question every NPS team fears: "So what do we do about it?"
The Aggregate Illusion is the moment a rich, disaggregated loyalty signal becomes a single dashboard metric. The score is technically accurate. It is also nearly useless as a decision tool because nothing in the calculation tells you which segment drives the score down, which touchpoint creates detractors, or which cohort is about to churn. Companies with consistently high NPS do not have a better formula — they have a better architecture: continuous collection at specific touchpoints, persistent stakeholder IDs linking every response to a known record, segment structure built at collection, and closed-loop action within days. Learn how this connects to program evaluation workflows and impact measurement design.
Most dedicated NPS tools — Delighted, AskNicely, SurveyMonkey's NPS feature, Medallia at the enterprise tier — were designed around the anonymous, quarterly, consumer-CX use case. They are genuinely effective when that is the actual need: transactional NPS for a retail brand with no requirement to contact detractors by name. Where they break down is the moment your program requires identity, disaggregation, open-ended theme extraction, or integration with a broader measurement workflow — beneficiary NPS, employee NPS, investee portfolio NPS, or any context where action on specific responses matters more than the aggregate trend.
Sopact Sense is a data-collection origin system, not an NPS-specific tool. Every respondent receives a unique stakeholder ID at first contact — whether that is a customer, a program participant, or an employee. That ID persists across every subsequent touchpoint, so disaggregation by tier, cohort, demographic, or touchpoint is automatic rather than retrofitted from an export. Qualitative "why" responses flow through the same record as the numeric rating, and Intelligent Column theme extraction produces coded themes by segment in minutes rather than the three to four weeks of manual coding that traditional NPS programs absorb before they can act.
To calculate NPS in Excel: put your 0–10 scores in column A, then use =COUNTIF(A:A,">=9")/COUNTA(A:A)*100 for Promoter %, =COUNTIF(A:A,"<=6")/COUNTA(A:A)*100 for Detractor %, and subtract the second from the first. The result is your NPS. For Google Sheets, the identical formulas work — no Excel-specific functions are required.
For segment-level analysis, use a pivot table with the segment field as rows and three calculated columns: Promoter count, Detractor count, and NPS. Excel-based NPS breaks down at two points: first, when response volume exceeds a few thousand rows, pivot recalculation becomes slow; second, when you need to merge open-ended text with score segmentation, Excel has no native qualitative analysis — you end up copying responses into a separate document for manual coding. That manual coding is the 3–4 week delay that makes most quarterly NPS programs unactionable.
Employee NPS uses the identical formula — eNPS = % Promoters − % Detractors — applied to the question "How likely are you to recommend [Organization] as a place to work?" on a 0–10 scale. Promoters (9–10), Passives (7–8), and Detractors (0–6) follow the same bands. The interpretation differs: eNPS typically runs lower than customer NPS because employees are tougher graders of their own workplace, so a +20 eNPS is respectable and a +40 is strong.
The most valuable eNPS segmentation is by department, manager, and tenure band (0–12 months, 1–3 years, 3+ years). New-hire eNPS almost always runs 15–25 points below tenured-staff eNPS, which exposes onboarding as a primary turnover driver — a signal that is invisible in the aggregate score. See pre-post survey design for the collection architecture that makes this visible in the first quarter, and training evaluation workflows for how eNPS integrates with workforce development programs.
The standard NPS scale is 0–10, not 1–5 — and the formula does not translate directly. Some survey tools offer "NPS on a 5-point scale" as a shortcut, but the resulting score is not comparable to a true NPS benchmark because the distribution of responses and the Promoter/Passive/Detractor bands are fundamentally different.
If you are locked into a 1–5 scale for legacy reasons, the closest approximation treats 5 as Promoter, 4 as Passive, and 1–3 as Detractor, then applies the standard subtraction. The number you get is internally comparable over time — you can see your own trend — but it cannot be compared to published 0–10 NPS benchmarks without explicit conversion, and vendors who claim otherwise are selling a false equivalence.
[embed: video]
The NPS formula is % Promoters minus % Detractors. Promoters score 9–10, Detractors score 0–6, and Passives (7–8) are counted in the response base but excluded from the subtraction. The result ranges from −100 to +100 and is reported as a whole number, not a percentage.
Calculate NPS by dividing the number of Promoters (scores 9–10) by total respondents, doing the same for Detractors (scores 0–6), and subtracting the Detractor percentage from the Promoter percentage. If 60% of 200 respondents are Promoters and 20% are Detractors, your NPS is +40.
Any NPS above 0 is positive — more respondents would recommend you than warn others away. Scores above +50 are excellent. Scores above +70 are world-class. Industry context shifts the baseline: a +20 in cable is strong; a +20 in e-commerce signals a problem. Your own trend over time is the benchmark that matters.
Passives (scores 7–8) count toward the total response base but are never included in the subtraction. If you have 100 respondents — 50 Promoters, 30 Passives, 20 Detractors — Promoter % is 50%, Detractor % is 20%, and NPS is +30. The Passive count affects your denominator but never your numerator.
The Aggregate Illusion is the moment NPS is calculated as one number across all customers, touchpoints, and time periods — collapsing segment-level signal into a single integer that describes everything and changes nothing. A +42 average can hide an onboarding NPS of +8, the real churn driver. The formula is trivial; escaping the Aggregate Illusion requires a system.
In Excel, use =COUNTIF(A:A,">=9")/COUNTA(A:A)*100 for Promoter %, =COUNTIF(A:A,"<=6")/COUNTA(A:A)*100 for Detractor %, and subtract the second from the first. For segment-level cuts, build a pivot table with the segment as rows and Promoter/Detractor counts as calculated columns.
Employee NPS uses the same formula — % Promoters − % Detractors — applied to the question "How likely are you to recommend [Organization] as a place to work?" on a 0–10 scale. eNPS typically runs lower than customer NPS, so +20 is respectable and +40 is strong. Segment by department, manager, and tenure to see what the aggregate hides.
Below 50 responses, NPS can shift 10+ points from statistical noise alone. For a confidence interval of ±5 points, aim for 200+ responses per segment. For smaller cohorts, report the trend across multiple survey cycles rather than treating any single score as a precise measurement — the direction of movement is more reliable than the absolute number.
Standard NPS uses a 0–10 scale with Promoters at 9–10, Passives at 7–8, and Detractors at 0–6. The 1–5 scale is a different instrument — the response distribution and band thresholds are not equivalent, so a "1–5 NPS" cannot be directly compared to published 0–10 benchmarks. Use 0–10 unless a legacy system prevents it.
Relational (annual) NPS has been largely replaced by transactional NPS tied to specific moments — onboarding completion, support ticket closure, renewal, program milestones. Transactional NPS runs continuously and produces actionable signal within days, not twelve months. Aggregating transactional scores across touchpoints into a single annual number re-introduces The Aggregate Illusion.
Dedicated NPS tools range from free tiers (Google Forms, basic SurveyMonkey) through $200–$3,000/month (Delighted, AskNicely) to enterprise contracts ($30K–$150K/year for Qualtrics or Medallia). Sopact Sense starts at $1,000/month and includes persistent stakeholder IDs, qualitative theme extraction, and segment architecture that dedicated NPS tools treat as add-ons or miss entirely.
Yes — Sopact Sense runs customer, beneficiary, and employee NPS programs with structural advantages dedicated NPS tools lack: persistent stakeholder IDs linking every response to a known record, disaggregation built at collection rather than retrofitted, and qualitative theme extraction that reads open-ended responses in minutes rather than weeks.
