What is Net Promoter Score? Calculation & Survey Tips

Net Promoter Score (NPS) is one of the most widely used metrics to measure customer loyalty. It captures how likely your customers are to recommend your brand to others, and, by extension, how strongly they believe in what you offer. That's the signal Fred Reichheld and Bain & Company set out to capture in 2003, and it's the reason roughly two-thirds of Fortune 1000 companies still run an NPS program of some kind.

The metric has aged well, but the way teams use it has not. Most NPS programs collect a number, drop it into a board slide, and move on. The companies that grow faster than their peers (and Bain's research shows loyalty leaders grow roughly 2x their competitors) treat NPS as a feedback loop, not a report.

This guide walks through what NPS actually measures, how to calculate it correctly, which survey type to run when, and the best practices that hold up in 2026.

What is Net Promoter Score?

Net Promoter Score (NPS) is a customer loyalty metric that measures how likely customers are to recommend a company, product, or service to a friend or colleague on a 0-10 scale. Developed by Fred Reichheld at Bain & Company in 2003, it segments respondents into Promoters (9-10), Passives (7-8), and Detractors (0-6).

* Promoters are enthusiastic supporters who are likely to recommend your brand. Passives are generally satisfied but may switch to competitors if a better option becomes available. Detractors are unhappy customers who may stop buying from your business and potentially discourage others from doing so.

It gauges overall customer sentiment towards a brand based on a single question: "On a scale of 0 to 10, how likely are you to recommend our company, product, or service to a friend or colleague?", resulting in scores ranging from -100 to +100. You can deploy your NPS survey after specific events (at checkout or website exit) to gauge how consumers feel about your brand at the moment.

The reason it's stuck around is straightforward: recommending a brand puts the customer's own reputation on the line, which makes the answer harder to fake than satisfaction. Bain's research consistently shows that NPS leaders grow at roughly twice the rate of their direct competitors, and that NPS explains 20% to 60% of the variation in organic growth between rivals in the same category.

How to calculate net promoter score (NPS)?

The formula is simple, which is part of why the metric spread so fast.

Walk through it in four steps:

1. Ask the standard question: "On a scale of 0 to 10, how likely are you to recommend [company/product] to a friend or colleague?"

2. Bucket the responses: Promoters (9-10), Passives (7-8), Detractors (0-6).

3. Convert each bucket to a percentage of total responses.

4. Subtract the Detractor percentage from the Promoter percentage. Passives stay in the denominator but don't get added or subtracted.

A quick worked example: You collect 200 responses. 100 score you 9 or 10, 60 score you 7 or 8, and 40 score you 0 to 6. Promoters are 50%. Detractors are 20%. Your NPS is 30.

The result lands somewhere between -100 (every respondent is a detractor) and +100 (every respondent is a promoter). Most companies sit between 0 and 50.

How to interpret your score

Absolute scoring thresholds are useful, but only as a sanity check. Use them, then compare them against your industry.

Source

The cross-industry average sits between 32 and 44 in 2026. Insurance and professional services lead. Telecom and ISPs sit at the bottom. SaaS hovers in the 30 to 40 range, and ecommerce around 45.

Why conducting an NPS survey is important

NPS is not the only loyalty metric you can run, but it's the one that comes closest to a leading indicator of revenue. Three reasons it earns its place in most CX programs:

• It links sentiment to growth. Bain's field research connects a 12-point NPS increase to a doubling of growth rate relative to competitors and shows that promoters generate roughly 2.6x the lifetime revenue of detractors. That makes NPS one of the few CX metrics finance teams will actually look at.
• It surfaces churn risk early. Around half of detractors churn within 90 days of leaving a low score. A clean NPS program with a detractor follow-up workflow gives customer success teams a 90-day window to recover the account before it lapses.
• It creates a shared language across the org. Product, support, marketing, and sales rarely agree on customer experience metrics. A relational NPS gives every team a single score to align around, while the verbatim feedback gives each function its own actionable list.

What are the different types of NPS surveys?

There are two primary types of NPS surveys, and they answer different business questions. Most mature programs run both.

1. Relational NPS (rNPS)

Relational NPS measures the overall health of the customer relationship. You send it on a fixed cadence, typically quarterly or semi-annually, to the entire customer base or a representative sample. The output is a loyalty trendline you can benchmark against past quarters, against peer companies, or against your annual targets.

Use rNPS when you want to:

• Track loyalty over time
• Report at a board or executive level
• Benchmark against industry averages
• Tie a single number to strategic CX initiatives

2. Transactional NPS (tNPS)

Transactional NPS measures sentiment after a specific event, sent within 24-48 hours of the interaction, so the experience is still fresh. The score reflects the touchpoint, not the relationship. Common trigger events include:

• Post-purchase
• After a support resolution
• After onboarding completion
• After a key product milestone (first integration, first report, first renewal)

Use tNPS when you want to isolate friction at a particular point in the journey. If your post-purchase tNPS is 60 but your renewal tNPS is 25, something specific is breaking down between those moments, and the verbatims will tell you what.

Example: Charles Schwab moved its NPS from -35 to +52 over 22 years, largely by running transactional NPS at the customer touchpoints where trust was being lost (account opening, fee disputes, advisor interactions) and rebuilding each one. The relational score followed the operational changes, not the other way around.

Where NPS surveys actually live

Beyond rNPS and tNPS, it's worth separating the channels you can run the survey through, because they each have different response-rate and bias profiles:

• Email NPS: Highest reach, average response rate of 15% to 30%. Strong for relational surveys.
• In-app or web NPS: Higher response rates because the user is already engaged. Best for transactional surveys triggered by usage events.
• In-product chat or messaging NPS: Conversational surveys consistently outperform static forms on completion and verbatim quality.
• SMS NPS: Useful for high-volume B2C contexts (retail, hospitality, telco) where email open rates are weak.
• Post-call IVR or voice NPS: Standard in contact-center environments, captures sentiment immediately after support interactions.

💡 Good to know: NPS methodology has also been adapted to measure employee loyalty as eNPS (Employee Net Promoter Score). eNPS uses the same 0-10 scale but asks employees how likely they are to recommend the company as a place to work. It's a separate program with different audiences and data sensitivity and is typically owned by HR rather than CX teams.

How to conduct an NPS survey

Running an NPS survey well is mostly about discipline. Five steps separate the programs that drive action from the ones that produce a slide deck.

Step 1: Define the goal before you pick the channel

Decide whether you're measuring the relationship or a specific moment. The answer determines the cadence, the audience, and the follow-up question. A relational NPS to your full customer base in October is a different program from a transactional NPS triggered after every support resolution.

Step 2: Use the standard question and standard scale

The wording matters. The 0 to 10 scale matters. Both are what make your score comparable to industry benchmarks. Customizing either breaks comparability and quietly biases your results. Stick with: "On a scale of 0 to 10, how likely are you to recommend [company/product] to a friend or colleague?"

Step 3: Add exactly one follow-up question

This is the single most important upgrade most programs need. Add an open-ended driver question right after the score: "What's the main reason for your rating?" or "What would make you more likely to recommend us?"

Driver responses are where you get the qualitative reasons behind the number. They're also what makes NPS feedback usable by product, support, and CX teams. Without them, you're tracking a number with no instructions.

Step 4: Close the loop fast

Closed-loop follow-up is the dividing line between a measurement program and a loyalty program. Within 24 to 48 hours of a low score, someone senior should reach out to the detractor, understand the issue, and resolve it if they can. Customers whose issues are resolved quickly often end up more loyal than customers who never had a problem in the first place.

Build the workflow once and automate it. Score thresholds trigger alerts, alerts route to the right account owner, and the resolution gets logged against the customer's record.

This is where modern CFM tooling like Sprinklr Surveys earns its place — conversational survey formats adapt follow-up questions based on the initial score, and detractor responses can auto-create cases routed to the right team with full context attached.

GRAB A DEMO NOW

Step 5: Analyze beyond the score

Three things to pull out of every NPS cycle:

• Driver analysis: Cluster verbatim responses to identify the top three reasons promoters love you, and the top three reasons detractors don't.
• Segment cuts: Score by customer segment (industry, plan tier, region, tenure). Aggregate numbers hide the most actionable patterns.
• Trend lines: Quarter-over-quarter and year-over-year movement matters more than any single number.

This is where AI-powered analysis earns its keep. Modern platforms like Sprinklr Surveys can run conversational NPS surveys that adapt follow-up questions based on the initial score, then auto-classify open-text responses into themes and root-cause drivers without a manual taxonomy. The output is a working list of what to fix, not a spreadsheet of comments to read.

Common pitfalls to watch for

A short list of mistakes that quietly inflate or deflate scores:

• Surveying only happy or recently purchased customers
• Letting sales or support pressure customers for a 9 or 10
• Running NPS during a major outage or pricing change without a control group
• Treating a 7 as a "pass" (passives are not promoters)
• Ignoring non-respondents (in B2B, silence is often a detractor signal)

Best practices to improve NPS survey response rate

A 15% to 30% email response rate is the industry baseline, and most of that range is leaking signal. The practices below consistently lift response rates and improve the quality of the data.

Keep the survey short

One question plus one follow-up. Anything longer is a different survey. Adding five satisfaction questions after the NPS scale meaningfully drops completion, and biases who finishes.

Time it around a meaningful moment

For transactional NPS, send within 24-48 hours of the event. For relational NPS, avoid spikes (renewal week, holiday peaks, a product launch) that color the response. Customers who answer in the wake of a single bad day will skew your number.

Personalize the invitation

A three-sentence email from a recognizable sender outperforms a templated, brand-only email by a wide margin. Use the customer's first name. Reference the touchpoint. Get to the question in the first line.

Make it mobile-first

Most respondents open survey emails on a phone. Buttons need to be tappable. The survey needs to load in under two seconds. The full experience (score, follow-up, thank-you) should be possible without leaving the inbox or scrolling more than once.

Send a single reminder

A reminder three to seven days after the original invitation typically lifts response rates by 20% to 40% without irritating respondents. Anything beyond one reminder hits diminishing returns and starts to feel like nagging.

Pick the right channel for the audience

Email is the workhorse for relational surveys. In-app or in-product surveys win for transactional moments because the user is already engaged. SMS is strong for B2C contexts where email open rates are weak. Match the channel to where your customer is at the moment you ask.

Respect survey frequency

Survey fatigue is real. About 74% of respondents will only answer up to five questions in a survey, and most customers should not see an NPS request more than once a quarter for relational and once per major touchpoint for transactional.

Where conversational survey formats help: Most response-rate practices above (short forms, adaptive follow-ups, contextual timing) get easier with conversational surveys that adjust as the customer responds. Sprinklr's Omnichannel Survey Software supports three delivery modes for the same NPS survey: standalone links that open as their own page, conversational surveys that run inside a chatbot on messaging channels, and proactive survey prompts that surface on websites and apps based on visitor behavior.

NPS vs CSAT vs CES

NPS measures loyalty. CSAT (Customer Satisfaction Score) measures satisfaction with a specific touchpoint. CES (Customer Effort Score) measures how easy a specific interaction was. They overlap, but they're not interchangeable. The table below pulls the practical differences together.

Quick rule of thumb: Use NPS for the relationship. Use CSAT for the touchpoint. Use CES for the workflow. Run all three if you have the operational maturity to act on each one, because each surfaces something the others miss. This is the operational case for a unified voice of customer methodology and a single platform behind it. Sprinklr Surveys is built on this premise: survey responses sit alongside data from social and digital channels, review sites, and service interactions, so insights from one source can be validated against others instead of read in isolation.

Final Thoughts

NPS works when it's treated as a system, not a score. The companies that consistently outperform their peers share a few habits: they ask the standard question, they pair it with open-ended follow-up, they route detractors to a real human within 48 hours, and they segment the data closely enough to act on it. The number on the dashboard matters far less than the operational rhythm behind it.

The other shift worth flagging for 2026 is the move from static, form-based surveys to AI-powered conversational ones that integrate with the rest of your voice-of-customer program. When NPS scores sit alongside social conversations, support transcripts, reviews, and product usage signals in one place, the score stops being a quarterly snapshot and starts becoming a continuous health metric.

Platforms like Sprinklr Insights consolidate solicited and unsolicited feedback in a single environment, which is where most enterprise programs are headed. Wherever you run the program, the discipline is the same: ask cleanly, act fast, and let the verbatims do most of the talking.