How to Design an NPS Survey: Audience, Questions, Responses, and Follow-Up
Published:
Last Updated:
Category: Marketing Glossary
Authors: Shusaku Yosa
An NPS survey needs more than a rating question. Decide whose experience you are investigating, what decision the results will support, and who will follow up. Changes to audience, wording, or timing can change what the score means.
This guide covers survey operations, from eligibility through follow-up. Use the NPS calculation guide for category counts and arithmetic. Recommendation intent is distinct from actual referrals, purchases, or retention.
Choose a relationship or an experience question
A survey about an overall relationship differs from one following onboarding, a purchase, or support. Bain distinguishes these NPS uses. They share a scoring format but should not be presented as one interchangeable trend.
Purpose | Example audience and timing | Decision supported |
|---|---|---|
Overall relationship | Eligible customers invited at a defined point | Investigate reasons by tenure or customer type |
Onboarding experience | A defined period after onboarding completion | Review explanation and initial-use difficulties |
Store pickup experience | Customers meeting the same completed-pickup condition | Investigate directions and handover problems |
Specify the response unit. One representative of a business account does not necessarily represent every user at that company. If customers who cancelled or failed to complete a process are excluded, document the exclusion and the experience the survey cannot describe.
Use a recommendation question and an optional reason
The following wording is a starting point. Replace the service name and test whether a small group understands the intended experience before launching. AAPOR's survey guidance supports clear, neutral questions that avoid combining several concepts.
Question | Suggested wording | Design check |
|---|---|---|
Recommendation | How likely are you to recommend this service to a friend or colleague? Choose 0–10. | Label 0 as not at all likely and 10 as extremely likely; do not preselect a score |
Optional reason | What is the main reason for your score? | Allow the person to continue without answering |
Optional follow-up permission | May a team member contact you to clarify your feedback? | Explain the purpose and contact method; allow refusal without disadvantage |
Avoid invitations that imply the expected answer, such as asking satisfied customers to select 9 or 10. If satisfaction is relevant, ask it separately. Under the standard NPS categories, 9–10 are promoters, 7–8 passives, and 0–6 detractors. Passives are included in all valid responses used as the denominator.
Agree on collection and information handling
Choose frequency around the customer experience and response burden. There is no universally appropriate monthly or quarterly schedule. Prevent overlapping invitations to the same person. Do not let account owners invite only the people they expect to be happy.
Design field | Record before launch |
|---|---|
Purpose and eligibility | Decision, date range, experience conditions, person or account unit |
Invitation and deadline | Send date, close date, reminder rules, delivery failures |
Validity and duplicates | Responses allowed per unit, which duplicate to retain, missing and out-of-range rules |
Information handling | Purpose, authorized access, retention approach, contact point |
Follow-up | Owner, urgent escalation route, deadline, completion record |
Comparability | Changes in wording, language, channel, eligibility, and timing |
An anonymous survey and an identifiable survey with individual follow-up are different arrangements. Explain who can identify respondents. Do not promise anonymity while routing named answers to sales. Ask people to avoid including unnecessary personal or third-party information in free text, and collect only what the investigation needs.
Inspect the response composition before the score
In a fictional example, all 500 eligible customers who completed onboarding receive an invitation and 100 provide valid ratings. The operational ratio of valid responses to invitations is 100 ÷ 500 = 20%. The opinions of the remaining 400 are unknown; the respondents cannot automatically be treated as all 500 customers.
Label this simple operational ratio clearly. Formal survey response-rate definitions may distinguish eligibility, delivery, and other dispositions. As AAPOR explains, the rate alone cannot establish survey quality or nonresponse bias.
- Record eligible audience, invitations, failed deliveries, responses, valid ratings, and exclusions.
- Compare respondent tenure, account type, and invitation channel with the eligible audience.
- Reconcile promoter, passive, and detractor counts to all valid ratings.
- Keep ratings valid when an optional reason is blank.
- Match comparison conditions and avoid excessive segmentation of small groups.
Assign follow-up based on issues, not scores alone
A low score warrants investigation, but a high score does not make a serious problem harmless. Safety concerns, billing problems, and major service failures should follow the agreed escalation route regardless of rating. Treat categories as a way to organize feedback, not as a diagnosis of the person or a promise that passives will become promoters after a small change.
The following fictional worksheet illustrates ownership, not demonstrated improvement.
Reported issue | Investigation or action | Owner and closure |
|---|---|---|
Setup instructions are difficult to find | Check the actual screen and route; consider revising the explanation | Onboarding owner records the change and verification |
Pickup point is unclear | Inspect the store directions and physical route | Store owner records verification and relevant inquiries |
Major service problem | Escalate without waiting for the survey report | Support lead records status, contact permission, and resolution |
Where respondents permit contact, explain what was investigated and done. Do not ask them to change their rating or make a benefit conditional on a future high score. After a change, examine recurrence, inquiries, retention, and purchases separately from NPS. The LTV evaluation guide and OMO experience design guide can help connect customer feedback to operational questions.
To organize resulting initiatives, owners, and review dates, explore Xtrategy's features.
