NPS question, how to word it and what to ask next

The classic wording is simple.

• How likely are you to recommend [company or product] to a friend or colleague?

Respondents answer on a scale from 0 to 10, where 0 means not at all likely and 10 means extremely likely. Keep the wording close to this standard if you want to compare your score over time. Small changes, such as "would you recommend" instead of "how likely are you to recommend", change how people read the question.

For business products, "to a colleague" often fits better than "to a friend". For consumer products, "to a friend or family member" can feel more natural. Pick one version and keep it.

Answers fall into three groups.

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

The Net Promoter Score is the percentage of promoters minus the percentage of detractors. Passives count toward the total number of responses but not toward either group. The result is a number between minus 100 and plus 100.

A quick worked case: if 100 people answer, 50 give 9 or 10, 30 give 7 or 8 and 20 give 0 to 6, your score is 50 minus 20, which is 30. The score alone tells you where you stand. It does not tell you why, which is the job of the follow-up.

Always add at least one open follow-up. The most common one is a single question for everyone.

• What is the main reason for your score?

You can also tailor the follow-up to the group the person falls into.

• For promoters: What do you value most about [product]?
• For promoters: What would you tell a colleague about us?
• For passives: What would make you more likely to recommend us?
• For passives: What is missing that would make this a 10?
• For detractors: What disappointed you?
• For detractors: What is the one thing we should fix first?

A short closing question helps you plan the next step.

• May we contact you about your answer?

Keep the whole NPS survey short: the score, one or two follow-ups and an optional contact question. Every extra item takes attention away from the follow-up, which is where the reasons live.

There are two common ways to run NPS, and they answer different questions.

Relationship NPS goes out on a regular schedule, such as every quarter or twice a year, to your whole customer base or a rotating sample. It measures how people feel about the company overall and shows trends over time.

Transactional NPS goes out after a specific moment, such as onboarding, a support case, a renewal or a delivery. It tells you how that moment affected loyalty.

Some timing rules apply to both.

• Wait until people have used the product enough to have an opinion.
• Do not send it right after a known outage or a billing problem unless that is what you want to measure.
• Avoid asking the same person more often than every few months.
• Send at a time when people read messages, not late on a Friday.

If you run NPS inside your product or on your site, website survey tools let you show it as a side tab or popup to logged in customers at the right moment.

A typed follow-up often gets a few words, such as "good service" or "too expensive". A spoken answer usually carries the story behind the score: what happened, when, and how it felt. To get that, word the follow-up as an invitation to talk.

• "In a minute or so, tell us what is behind your score. What has worked for you, and what has not?"
• "Think about the last time you used [product]. What happened that made you give this score?"
• "If a colleague asked you about us today, what would you actually say?"

Let people choose the format. Some will record a voice note on the way to a meeting, some will record a short video at their desk and some will type. With voice feedback and video surveys, the same question accepts all three, and spoken answers are transcribed so you can read them like text.

A score means most when you compare it with your own past results asked the same way. Compare like with like: the same question wording, the same scale, the same type of customer and the same moment in the journey. A relationship score sent to all customers and a transactional score sent after support tickets measure different things and should not sit in one trend line.

Be careful with small groups. When only a handful of people answer, one or two responses move the score a lot. Report the number of responses next to every score so readers can judge how much weight to give it.

Published benchmarks from other companies vary by industry, region, survey method and timing, so treat them as loose context at best. Your own trend, split by segment, is a far more reliable guide to whether things are getting better.

NPS is easy to run and easy to run badly. These are the most common problems.

• Changing the scale. A 1 to 10 or 1 to 5 scale is not NPS and cannot be compared with NPS benchmarks or your own history.
• Coloring the scale. Red, yellow and green numbers nudge people toward certain answers. Use a neutral scale.
• Skipping the follow-up. A score without reasons gives you a number to report and nothing to act on.
• Asking for a high score. "We would love a 10" in the email changes the answers. Ask neutrally.
• Surveying only happy customers. If you send NPS only after successful moments, the score looks better than reality.
• Treating the score as the goal. Teams that chase the number instead of the reasons tend to stop learning.
• Never closing the loop. If people take time to explain a low score and hear nothing back, they stop answering.

Start with the score and its trend, then split it by segment: plan, region, account size, tenure or product area. A flat overall score can hide one segment improving and another falling.

Then read the follow-ups, starting with detractors. Tag each answer with themes such as onboarding, reliability, support, pricing or missing features. Count how often each theme appears in each group. Themes that show up often among detractors are your fix list. Themes that show up often among promoters are what to protect and what to say in your marketing.

Look at passives closely. They are often one fix away from becoming promoters, and their answers tend to name that fix directly.

Pull a few quotes per theme and share them with the teams who own each area. A short clip or quote from a real customer travels further inside a company than a chart. Finally, decide on two or three actions, assign owners and tell respondents what you changed.

When answers pass a few dozen, tagging by hand gets slow. Survey analysis software transcribes spoken answers, groups them into themes, marks sentiment per answer and pulls quotes with timestamps so the reading part takes minutes instead of days.

Closing the loop means replying to people after they answer. For detractors, a short personal reply that acknowledges the problem and says what happens next goes a long way. For promoters, a thank you and, where it fits, an invitation to share a review or a short video testimonial software recording turns goodwill into proof other buyers can see.

Keep a simple rule: anyone who left a contact permission and a score of 6 or below gets a reply within a few working days. Track how many of those replies go out. It is one of the few NPS metrics fully in your control.