Sample NPS Survey Email Template: Copy-Paste Variants
Key takeaways
- Send your NPS survey within the memory window of the event: 24-48 hours post-purchase, 3-7 days post-onboarding, or within hours of a support ticket.
- Keep the subject line under six words, drop the word "survey," and send from a real person to maximize open rates.
- Embed the 0-10 scale directly in the email body using plain HTML anchor links to eliminate landing page redirects and login walls.
- Maintain identical question wording across all send contexts to keep your trend line comparable, personalizing only the setup sentence.
- Follow up within 24 hours: ask promoters for a review, ask passives what went sideways, and offer detractors a direct fix or refund.
The sample NPS survey email template below is ready to paste into Gmail or Outlook today, with four variants matched to when you send: post-purchase, post-onboarding, post-support, and post-event. Each subject line, body line, and CTA is annotated with the copy decision behind it, plus a plain-HTML fallback for clients where embedded surveys break.
When should you send an NPS survey email relative to the customer's action?
Send within the memory window of the event you're measuring. Post-purchase: 24-48 hours after delivery, once the product is in hand but the unboxing is fresh. Post-onboarding: 3-7 days after signup, after the customer has actually used the core feature, not the day they registered. Post-support: within a few hours of ticket resolution, while the interaction is still vivid. Post-event or newsletter: same day or next morning.
Send too early and they can't judge the product; send too late and response rate decays along with recall. Timing matters even more if you embed the question in the email itself, because interactive forms render inconsistently across clients, as the Can I Email support tables document for Gmail and Outlook. One Question handles that rendering problem for you; its pricing includes a solo tier if you send from Gmail or Outlook yourself.
What should the subject line of an NPS survey email say?
Keep it short, name the sender, and drop the word "survey" entirely. "Survey" reads as homework; a one-click question reads as a favor you can answer in two seconds.
Subject lines that read as a person, not a program
Annotate each send context like this:
- Post-purchase: "Quick question about your order, Sarah?" — first name plus "quick" signals one-tap effort, and it sounds like the shop owner, not a ticketing system.
- Post-onboarding: "How's it going so far?" — conversational, no ask stated, so opening feels low-commitment.
- Post-support ticket: "Did we fix it?" — four words, ties directly to the resolution they just experienced.
- Post-event or newsletter: "Worth your time?" — asks for a verdict, which maps cleanly to a 0–10 score.
Pair every subject line with a human From name ("Ana at [Shop]"), since sender recognition drives opens more than clever copy. One Question fits this pattern: you send from your own Gmail or Outlook, so the From name is already a real person.
What does a sample NPS survey email template look like for post-purchase sends?
Copy this into your email body and swap the bracketed fields:
Subject: Quick question about your [product name] order
Hi [first name],
Your order arrived on [date]. Based on your experience so far, how likely are you to recommend [brand] to a friend?
0 — Not at all likely | 10 — Extremely likely
One tap, no login. Your reply goes straight to [your name].
[Your name], [brand]
Line-by-line: why each part of the template is written that way
The subject line names the order, not the survey, so it reads like a check-in rather than a request. The body opens with the delivery date to anchor the score in a real event. The question sits in the email itself: every extra click and page load costs respondents, which is why shops that keep the question in-email report better completion than link-out versions of the same survey (see Survicate's in-email surveys for the same pattern). The "one tap, no login" line removes the two biggest silent drop-off causes. Sign it with a person's name, not the brand, so replies land in a monitored inbox.
How do you adapt one template for post-onboarding versus post-support sends?
The skeleton stays the same: subject line, one-line setup, the 0–10 question embedded in the email, one-tap scoring. What changes is the setup sentence, the send timing, and the follow-up promise.
Post-onboarding variant
Send it 3–7 days after signup, once the customer has actually used the product. Swap the setup line to reference the outcome: "You've been set up for a few days now — quick check-in." Drop the delivery or order reference entirely. If they score low, your follow-up should offer a live call, not a ticket link, because an onboarding detractor usually needs hands-on help rather than a fix.
Post-support ticket variant
Send it within an hour of ticket closure, while the interaction is fresh. The setup line names the ticket: "Quick one-tap rating on how we handled your ticket." Keep the body to two sentences. A support NPS measures the person who helped, so resist adding a second question about the product — that split muddies the signal and adds a question you don't need.
The reason one skeleton works for both is that the respondent never reads the whole email — they read the subject line, the first sentence, and the score buttons. Everything else is friction. So when you adapt the template, change only those three elements and leave the rest untouched. A post-onboarding send that opens with "You've been live for a few days" and a post-support send that opens with "How did we do on your ticket?" feel like different emails to the recipient, but they share identical structure underneath.
That consistency is also what keeps your data comparable: same question, same scale, same one-tap buttons, asked at two different moments in the customer's life. If you rewrite the question wording per context ("How likely are you to recommend us to a fellow store owner?"), you break comparability between the two sends, and your trend line stops meaning anything. Resist the urge to personalize the scale itself — personalize the setup, never the question.
What should the body copy contain if you embed the question directly in the email?
Put the 0-10 scale in the email itself. Every click to a landing page, every page load, and every extra field costs you respondents, so the body copy needs exactly three things: the question, a one-tap row of 0-10 buttons, and nothing else above the fold.
Making the 0-10 scale one-tap and no-login
The tap should record the score immediately, with no login. One Question adds one Confirm tap on a short page, so clicks from email security scanners are kept apart from answers people actually gave. Gmail and Outlook strip JavaScript and much CSS, so build the scale as plain HTML anchor links or image buttons with alt text, and keep a text fallback ("reply with a number 0-10") for clients that break the markup.
| Approach | Clicks to respond | Survives Gmail/Outlook |
|---|---|---|
| Link-out survey | 2+ (click, load, answer) | Yes |
| Embedded one-tap HTML | 1 | Yes, with fallback |
| Embedded widget/JS | 1 | No, often stripped |
Dedicated tools handle the plumbing for you. Survicate and Alterable both offer in-email NPS surveys, and One Question works in a similar way: paste the question into a Gmail or Outlook email and the answer buttons show inline. It offers Faces, Yes / No, a 1–5 List and Open text rather than a 0–10 scale, and only confirmed answers are counted. If you paste your own HTML today, the one-tap row is the one upgrade worth copying.
How does an NPS survey email template differ from a CSAT survey email template?
NPS asks about the whole relationship ("how likely are you to recommend us?"), so it works best as a periodic check-in, like 90 days after purchase. CSAT asks about one interaction ("how did we do on this order?"), so it fires right after a specific event.
That changes the wording. NPS copy stays broad and relationship-focused; CSAT copy names the transaction, the ticket, or the delivery. Scale differs too: NPS uses the 0–10 recommend question, while CSAT usually runs a 1–5 or emoji rating.
Follow-up splits the same way. Detractors get a "what went wrong?" reply; low CSAT scores get a fix for that specific order or ticket. If you want both, send NPS quarterly and CSAT per transaction, and keep each email to its one question.
How do you follow up with promoters, passives, and detractors after the survey?
Segment by score: 9–10 are promoters, 7–8 passives, 0–6 detractors. Each band gets a different reply, sent within 24 hours while the purchase is still fresh. The scoring bands come from the original Net Promoter method described by its creators at Bain.
Promoter follow-up
Subject: Thanks for the 10 — one small ask
Body: Thanks for rating us a 10. If you have 30 seconds, a review helps other shoppers find us. Either way, thanks for buying.
One ask, one link, no login.
Passive follow-up
Subject: Quick check-in on your order
Body: You scored us a 7 or 8, which usually means something specific went sideways. Reply with one word for what it was — shipping, sizing, or something else.
Detractor follow-up
Subject: You rated us low — I read every reply
Body: Sorry we missed. Reply with what happened and I'll fix it or refund it, no forms, no back-and-forth.
Reply personally from your own inbox; a templated blast defeats the point.
What common copy mistakes depress NPS survey response rates?
Run this checklist before sending:
- Multi-question creep. One score, one optional comment. Every added question costs completions.
- Login walls. If responding needs an account, most people stop. Keep it one click, no login.
- Vague CTAs. "Share your feedback" loses to "How likely are you to recommend us? Click 0–10."
- Wrong sender. A no-reply address signals nobody reads answers. Send from a person.
- Over-long preamble. Two short sentences of context max; the question should be visible without scrolling.
Plain HTML buttons handle the reliability problem, but copy mistakes are yours to fix. If a send underperforms, change one variable at a time so you know what moved the response rate.