NPS survey template
You need a recommendation score you can track over time, and you need to know what to do about it — not just the number.
What this template is for
Most NPS surveys collect a number and stop there. A score of 32 tells you nothing about what to change. This template keeps the standard 0–10 question so the score stays comparable, then branches: detractors are asked what went wrong (required), promoters are asked what you do better than the alternatives, and everyone reports usage frequency so the score can be read per segment.
Questions
Every question below is part of the actual survey definition — the same one the preview and “Use this template” use.
How likely are you to recommend us to a friend or colleague?
Rating · 0–10 · Required
Why this question: Kept as the standard 0–10 wording so results stay comparable with your own history and with external benchmarks. Changing the scale or the wording breaks that comparability, which is the main reason to use NPS at all.
Watch out: Do not relabel the endpoints to something friendlier. "Extremely likely" is part of what makes the score comparable.
What is the main reason for your score?
Long text · up to 600 characters · Required
Shown when: Detractor
Why this question: Required only for scores 0–6. A detractor who leaves without explaining is the most expensive kind of response: you know something is wrong and cannot act on it.
Watch out: Making this required for everyone lowers completion; making it optional for detractors loses the responses you most need.
What do we do better than the alternatives?
Long text · up to 600 characters
Shown when: Promoter
Why this question: Promoters describe your actual differentiator in their own words. This is the raw material for positioning copy, and it is more honest than asking "why did you choose us" at signup.
What would make your experience clearly better?
Long text · up to 600 characters
Why this question: Passives (7–8) are the group most likely to move with a small change. Asking what would make the experience clearly better points at the cheapest wins.
How often do you use the product?
Single choice · Required
Why this question: Frequency is the segment that most often explains score differences. Without it, a falling score cannot be separated from a shift in who is answering.
Which best describes your role?
Single choice + Other
Why this question: Role separates "the person who uses it" from "the person who bought it". These two groups routinely score several points apart, and the gap is the interesting part.
How likely are you to recommend us?
How likely are you to recommend us to a friend or colleague?
What is the main reason for your score?
What do we do better than the alternatives?
What would make your experience clearly better?
How often do you use the product?
Which best describes your role?
When to use this
- You want one metric you can track quarter over quarter across the whole customer base.
- You already have enough respondents that segment-level differences are meaningful (roughly 100+ responses per segment).
- You want verbatim reasons attached to the score, not collected in a separate survey.
When not to use this
- You want to evaluate one specific interaction — a support ticket, a checkout, an onboarding call. Use a CSAT or CES survey; NPS measures the overall relationship and will be dominated by prior experience.
- You have fewer than about 50 responses. NPS is a difference of two proportions, so it moves violently at small n and you will chase noise.
- You need to know whether a specific feature works. Ask about the feature directly; a recommendation score cannot isolate it.
Branching
Page 2 shows a different follow-up per score band: 0–6 gets a required reason, 9–10 gets the differentiator question, and 7–8 sees the improvement question. Everyone answers usage frequency and role, so every response is segmentable regardless of which branch they took.
Validation
The score is required — an NPS response without a score is not a response. Free-text answers are capped at 600 characters, which is long enough for a real explanation and short enough that the responses stay readable in bulk.
What to look at once responses arrive
Read the distribution before the score
A single NPS number hides its own shape. Two products with the same score can have completely different distributions — one with everyone at 7–8, one split between 10s and 3s. The second is far more urgent.
Split by usage frequency
If daily users score lower than monthly users, the product gets worse the more you use it. That is a retention problem, not a marketing problem.
Read detractor verbatims as a group, not individually
Individual complaints feel urgent. Grouping them shows which cause actually repeats — usually three or four themes cover most of the volume.
Track the same question over time
The absolute number matters less than the direction. Keep the wording frozen so quarter-over-quarter changes mean something.
Crosstabs worth running
How often do you use the product? × How likely are you to recommend us to a friend or colleague?
Shows whether the score is driven by heavy users or light users. A low score concentrated in daily users points to a product problem; concentrated in rare users it usually points at onboarding.
Which best describes your role? × How likely are you to recommend us to a friend or colleague?
Buyers and users score differently. If executives score high and individual contributors score low, renewals look safe while daily satisfaction is eroding.
Charts worth building
bar — How likely are you to recommend us to a friend or colleague?
The raw 0–10 distribution. Look for the shape, not the average.
stacked-bar — How often do you use the product? × How likely are you to recommend us to a friend or colleague?
Promoter/passive/detractor mix per usage segment, so you can see which segment moves the score.
Customise it with AI
Paste this into SurveyZero’s assistant to adapt the template to your product. It keeps the parts that make the results comparable and changes the parts that should be specific to you.
Take this NPS survey and adapt it to a B2B analytics product sold to data teams. Keep the 0–10 recommendation question exactly as-is so the score stays comparable, keep the branching by score band, and replace the usage frequency and role options with ones that fit data teams. Add one question about which part of the workflow they rely on most.
Related templates
- CSAT survey template — You want to measure one specific interaction — usually a support ticket — right after it happens, and act on the bad ones the same week.
- Churn survey template — Customers are cancelling and you need to know which cancellations you could have prevented — not a list of complaints.