The NPS Survey in 2026: Questions, Timing, and the Follow-Up That Matters

What is an NPS survey?#

An NPS survey is a two-part customer feedback survey that measures loyalty by asking one rating question — "How likely are you to recommend us to a friend or colleague?" on a 0–10 scale — followed by one open-ended question asking why. The rating produces your Net Promoter Score; the follow-up question is where the reason behind the score actually lives.

That structure has barely changed since Bain & Company's Fred Reichheld introduced the metric in the 2003 Harvard Business Review article "The One Number You Need to Grow." What has changed in 2026 is how much value teams now expect from the second question — and how badly a single-line text box serves it. This guide covers the questions, the timing, the channels, and the open-ended follow-up that turns a number into something you can act on.

The two-question NPS survey structure#

Every net promoter survey is built from two questions: one closed rating question and one open-ended follow-up. The rating question is standardized so scores stay comparable across companies and over time; the follow-up is where you learn what the number means.

Question 1 — the rating (the standard wording):

On a scale of 0 to 10, how likely are you to recommend [company/product] to a friend or colleague?

Responses map to three segments: Promoters (9–10), Passives (7–8), and Detractors (0–6). Your score is the percentage of Promoters minus the percentage of Detractors, which lands somewhere between −100 and +100. If the segments or the math are new to you, start with the 0–10 NPS scale and its three segments and the step-by-step NPS formula with worked examples. For the metric's full definition and lineage, see what Net Promoter Score is and the why behind it.

Question 2 — the open-ended follow-up:

What's the primary reason for your score?

Question 1 is the metric. Question 2 is the diagnosis. Most teams treat the rating as the deliverable and the follow-up as an optional afterthought — which is exactly backwards, and the core argument of this guide.

Relational vs transactional NPS surveys#

The two NPS survey types differ by trigger: relational surveys run on a fixed schedule, while transactional surveys fire after a specific interaction. Choosing between them determines your timing, your wording, and your expected response rate.

Relational NPS answers "how do customers feel about us overall?" Transactional NPS answers "how did that specific moment land?" Mature programs run both, and map each to the relevant stage in the customer lifecycle and its metrics.

When and where to send an NPS survey#

Send transactional NPS surveys within hours of the interaction and relational NPS surveys on a quarterly cadence, using the channel where your customers already are. Timing and channel drive response rate more than any other design choice.

Timing by survey type:

• Transactional: trigger within 0–48 hours of the interaction while it's fresh. The sooner you ask, the more accurate and complete the memory you capture.
• Relational: run quarterly for B2B accounts and every six months for B2C. More frequent than that and you invite survey fatigue; less frequent and you miss shifts.
• Suppression: after a customer responds, suppress them from further NPS prompts for at least 60 days, and coordinate across teams so the same person isn't hit with NPS, onboarding, and CSAT requests in the same month.

Channels, ranked by depth vs reach:

1. In-app / in-product — highest response rates because you catch users in context; best for product-led relational NPS.
2. Email — the workhorse for relational surveys; broad reach but the lowest response rate of the three.
3. SMS / chat — high open rates and strong for transactional, post-interaction moments.

Commonly cited industry benchmarks put email NPS response rates in the 5–15% range, with transactional in-app and post-interaction surveys often reaching 20–40% because they ask at the moment of relevance. Response rate matters for representativeness, but it is not the same as insight — a 40% response rate of one-word comments still leaves you guessing. That distinction is the whole point of the next section, and it connects directly to the early-warning signals traditional surveys miss.

Writing the open-ended follow-up question#

The open-ended follow-up is the single most valuable line in the entire net promoter score survey, so write it to invite a specific reason rather than a generic complaint. A vague prompt gets vague answers; a targeted prompt gets root causes.

Weak follow-up (avoid): "Any additional comments?" — signals the survey is over and invites blank fields.

Strong follow-ups (use one, matched to the segment):

• All respondents: "What's the primary reason for the score you gave?"
• Detractors (0–6): "What was the most frustrating part of your experience, and what would have made it better?"
• Passives (7–8): "What's one thing that would have made this a 9 or 10?"
• Promoters (9–10): "What do you value most about [product], and who else would find it useful?"

Segment-specific wording is the highest-leverage upgrade you can make to an NPS survey question set. A promoter and a detractor are answering completely different questions in their heads, and asking them the same generic line wastes both. Routing the right follow-up to the right segment is a form of what we cover in customer feedback types and how to act on them — the follow-up is where an NPS survey stops being a metric and starts being genuine voice-of-customer input.

Why a conversation beats a comment box#

A conversation captures the reason behind a score far better than a comment box because it can ask a second question — the box cannot. This is the structural limitation of every static NPS survey, and it's why teams increasingly treat the open text field as a starting point rather than the finish line.

Think about what actually happens in a one-line follow-up. A detractor types "too expensive" and hits submit. That's the most common feedback teams collect and the least useful, because "too expensive" has at least four distinct meanings: the price is genuinely too high, the value isn't clear, a competitor is cheaper, or the customer hit a surprise charge. A comment box can't tell which. A follow-up question — "Compared to what? What would have made the price feel worth it?" — resolves it in one exchange. The static form has no way to ask.

This is the gap why traditional NPS surveys are not enough documents, and the reason product teams are rethinking how they run NPS in 2026. Even Reichheld has since argued in Bain & Company's "Net Promoter 3.0" that the score on its own gets gamed and manipulated, and that the qualitative "why" is what actually connects loyalty to growth. The fix isn't abandoning the metric — the 0–10 number is a useful, comparable trend line. The fix is replacing the dead text box with something that can follow up.

That's the approach behind the conversational NPS method that captures the why behind the score. Instead of a static comment field, an AI interviewer reads the score, asks a tailored follow-up, and probes vague answers the way a researcher would — at the scale of an automated survey. Perspective AI runs exactly this: hundreds of these short conversations at once, so the "why" arrives already structured, not as a spreadsheet of one-word fragments you have to code by hand. It's the same logic that makes conversations outperform surveys for real customer research, applied to the NPS moment specifically.

The rating still gives you the trend. The conversation gives you the roadmap.

NPS survey templates#

Use the templates below as ready-to-ship starting points, then adapt the follow-up wording to your segment and product. Each pairs the standardized rating question with a follow-up designed to surface a reason, not a shrug.

Relational NPS template (quarterly):

1. How likely are you to recommend [company] to a friend or colleague? (0–10)
2. What's the primary reason for your score?
3. (Optional, for 0–8) What's the single change that would most improve your experience?

Transactional NPS template (post-support):

1. Based on your recent support experience, how likely are you to recommend [company]? (0–10)
2. What worked well or fell short about how we handled your issue?
3. (For 0–6) What would have resolved this to your satisfaction?

Onboarding NPS template (post-activation):

1. After getting started with [product], how likely are you to recommend it? (0–10)
2. What was the hardest part of getting up and running?
3. (For 9–10) What clicked for you fastest?

For a fuller picture of how NPS sits alongside CSAT, CES, and CLV — and when to reach for each — see the 8 customer experience metrics that matter in 2026 and the deep dive on the CSAT formula, benchmarks, and limits. If you're weighing whether the whole survey-based model still fits, the shift from survey-based CX measurement to conversational VoC makes the broader case, and CX and product teams can see how this maps to their workflows on the pages built for CX teams and built for product teams.

Frequently Asked Questions#

What questions are on an NPS survey?#

An NPS survey has two questions: a 0–10 rating question ("How likely are you to recommend us to a friend or colleague?") and one open-ended follow-up asking the reason for the score. The rating produces the Net Promoter Score, while the follow-up explains it. Some programs add a short segment-specific prompt for detractors or promoters, but two questions is the standard and keeps completion rates high.

How long should an NPS survey be?#

An NPS survey should be as short as possible — ideally the one rating question plus one open-ended follow-up. The metric's appeal is its brevity, and adding extra questions measurably lowers completion rates. If you need more depth, capture it in the follow-up conversation rather than by stacking more closed questions, so you keep the survey short while still learning the why.

When is the best time to send an NPS survey?#

Send transactional NPS surveys within 0–48 hours of the interaction and relational NPS surveys on a quarterly cadence for B2B or every six months for B2C. For email, mid-week mornings in the recipient's local time zone tend to perform best. Always suppress a respondent from further NPS prompts for at least 60 days to avoid survey fatigue.

What is a good NPS survey response rate?#

A good NPS survey response rate depends on the channel and type: email relational surveys commonly land in the 5–15% range, while transactional in-app or post-interaction surveys often reach 20–40%. Higher is better for representativeness, but a high response rate of one-word answers still leaves the "why" unanswered — depth of response matters as much as volume. See what counts as a good NPS score for score benchmarks.

What is the difference between relational and transactional NPS surveys?#

Relational NPS surveys run on a fixed schedule and measure overall sentiment toward a brand, while transactional NPS surveys fire right after a specific interaction and measure sentiment about that touchpoint. Relational NPS answers "how do customers feel about us overall?"; transactional NPS answers "how did that moment land?" Mature programs run both and map each to the relevant lifecycle stage.

Do open-ended NPS follow-up questions actually get answered?#

Yes — open-ended follow-ups get meaningful answers when they're specific and asked at the right moment, but a static text box often collects short, ambiguous fragments like "too expensive." A conversational follow-up that can ask a clarifying second question resolves that ambiguity, which is why teams increasingly replace the comment box with a short AI-moderated exchange. Depth comes from the ability to follow up, not from the length of the survey.

Conclusion#

A well-designed NPS survey is still one of the simplest, most comparable ways to track customer loyalty: two questions, a 0–10 scale, and a cadence that fits how customers actually experience your product. Get the timing and channels right, keep it short, and route a segment-specific follow-up to each respondent, and you'll have both a reliable trend line and the beginnings of a reason behind it.

But the reason is where the value lives — and a single-line comment box can't ask "why" a second time. That's the one upgrade worth making in 2026: keep the rating question for the trend, and replace the dead text field with a conversation that can probe. Start a conversational study in minutes to see how AI-moderated NPS follow-ups turn "too expensive" into a specific, actionable reason — or browse example studies to see the format in action. Your NPS survey already tells you who your promoters and detractors are; the follow-up is how you find out what to do about it.