The Best NPS Survey Questions (And the One You Have to Ask)

NPS is the simplest customer-feedback survey in the world. One question. One follow-up. A score you can chart over time.

It's also the most commonly broken. Companies ask 8-question "NPS surveys" that aren't NPS, send them at the wrong cadence, ignore the follow-up text, and end up with a score that means nothing.

Here's exactly what to ask, how to ask it, and what to do with the answers.

The one question

The Net Promoter Score question is:

"How likely are you to recommend [company/product] to a friend or colleague?" — on a scale of 0 to 10.

That's it. That's the question.

• 9 or 10: Promoter
• 7 or 8: Passive
• 0 to 6: Detractor

Score formula: % Promoters minus % Detractors = NPS. Range is -100 to +100. Anything above 0 means more people would actively recommend than warn off. Above 30 is good. Above 50 is exceptional.

Do not modify the wording. Do not change the scale. Do not add adjectives. The whole point of NPS is comparability — your score should be comparable to industry benchmarks and to your own past data. Modifying the wording breaks both comparisons.

Anonymeter's 5-star adaptation

Anonymeter's rating UI is 5 stars (consistent with the rest of the platform). For NPS specifically, we adapt: 5★=Promoter, 4★=Passive, 1–3★=Detractor. Internally the trend works the same. For external industry benchmarking that requires the standard 0–10 scale, use a multiple-choice question with 11 options instead. See our NPS template →

The one follow-up question

The score is useless without context. Every NPS form should include:

"What's the main reason for your score?" — open text, optional

The text responses are where the value is. A 4-star score with no comment tells you nothing. A 4-star score with "I love the product but my CSM hasn't responded in 2 weeks" tells you exactly what to fix.

Don't make this required. Optional gets higher response rate; the people who do write something are the ones whose feedback is most useful.

The optional third question (sometimes worth it)

"What would make you more likely to recommend us?" — open text, optional

Forward-looking. Captures actionable improvement asks. Pair this with the score-reason question to get the highest-value answers — what they think + what they want.

Three questions is the upper limit. Don't add anything else. Every additional question drops your completion rate by ~10%.

What NOT to ask

These are the questions that turn a clean NPS survey into a bloated one:

• "How long have you been a customer?" — kills response rate. You have this data in your CRM anyway.
• "What's your role/title?" — same. Get it from your CRM.
• "What features do you use most?" — different survey. Don't bundle.
• "Rate the following: support, product, pricing, billing..." — that's a CSAT survey, not NPS. Run it separately or you're conflating two different metrics.
• "What's your email?" — if your goal is anonymous NPS for higher response rates, don't gate it. If your goal is attribution, send the survey to identified customers and skip the form-side email question.
• "Why?" before showing the score input — biases the response. Always show the score question first, then the follow-up.

When to send NPS

The two flavors:

Relationship NPS (the standard)

Sent on a regular cadence (quarterly is the industry standard) to your customer base. Measures overall sentiment about the relationship.

• Quarterly cadence is the sweet spot
• Survey ~25% of your base each cycle (so each customer gets one survey per year)
• Don't survey the same customer more than once per quarter — fatigue collapses response rates
• Avoid major events (right after a price increase, just after a known outage) — the score will reflect the moment, not the relationship

Transactional NPS (use sparingly)

Sent after a specific interaction (support ticket closed, onboarding completed, feature launched).

• Send within 24 hours of the trigger event
• Limit to 1 transactional NPS per customer per month (combine cadences if multiple events fire)
• Track these separately from relationship NPS — they measure different things

Most teams overdo transactional NPS and underdo relationship NPS. The relationship score is the strategic indicator; transactional NPS is the tactical signal.

How to maximize response rate

Standard email-gated NPS: 5–15% response rate. Anonymous embedded NPS (form on your dashboard or in-product): 25–40%.

The biggest single lever: embed the form, don't email it. Email-gated NPS skews toward your most engaged customers (Promoters). Embedded NPS captures the broader distribution including the Detractors who'd never log into a survey portal.

Other levers:

• No login wall. Login walls cut response rate by 30–50%.
• One question above the fold. The score should be the first and only thing the respondent sees before scrolling.
• 2–4 weeks between waves to the same customer. Closer = fatigue. Further = the moment has passed.
• Mobile-friendly. ~60% of NPS responses come from mobile. If your form is slow on mobile, you're losing data.

What a good NPS score looks like (by industry)

Wildly varies. Don't chase an absolute number — chase your trend over 4+ quarters.

Approximate industry benchmarks (2026):

• B2B SaaS: 30–50 is healthy
• Consumer SaaS: 20–40
• E-commerce / retail: 30–60
• Banking / financial services: 20–40
• Telecom / cable: -10 to +10 (yes, often negative)
• Healthcare: varies wildly (10–60)
• Government services: typically negative

If your industry benchmark is 30 and you score 60, you're doing exceptionally well. If your industry benchmark is 30 and you score 35, you're slightly above average — which is fine but not a moat.

What to do with the score

The biggest single mistake: collecting NPS and never sharing the text responses with the team.

The cycle that works:

1. Weekly: read the new text responses personally (5–15 minutes). Pattern matching is most useful when you do it manually for the first 200 responses.
2. Monthly: tag text responses by theme (pricing, support, specific feature, onboarding, etc.). The theme distribution among Detractors is your priority list.
3. Quarterly: report the score, the trend, and the 2–3 dominant themes to the whole company. NPS that's never shared is data that never drives action.
4. Annually: look back at whether the themes you addressed in past quarters shrank in volume. If they did, the program is working. If they didn't, dig into why.

Anonymous vs identified NPS

The trade-off:

• Anonymous embedded NPS: 3–4× higher response rate, more honest text, captures churning customers. Loses per-account attribution.
• Identified email-gated NPS: lower response rate, more polite text, biased toward Promoters. Gives per-account attribution for follow-up.

Most teams run both. Anonymous for the honest aggregate signal; identified for the customers you specifically want to talk to.

If you have to pick one, anonymous wins for most use cases. The honest text is more valuable than the per-account attribution.

Bottom line

NPS works because it's simple. Don't break that. One question, one optional follow-up, sent at the right cadence, with the score chart-able across years. Read every text response. Act on the themes.

Run an anonymous NPS survey in 5 minutes → — or see how NPS differs from CSAT and CES →.