BlogGuide

How to run an NPS survey inside your product

Net promoter score works best asked in context, to the right people, once. Here is how to set it up in a few minutes, and how to read what comes back.

The Respondo team · 3 min read

Net promoter score asks one question: how likely are you to recommend us to a friend or colleague? People answer on a scale from 0 to 10, and each answer puts them in one of three groups.

  • Promoters answer 9 or 10. They are likely to recommend you.
  • Passives answer 7 or 8. Satisfied, but not enthusiastic.
  • Detractors answer 0 to 6. Something is not working for them.

The score is the percentage of promoters minus the percentage of detractors. Passives count towards the total but not towards either side, so the score runs from −100 to +100.

The number is only as good as who you ask and when. An email a week later reaches the people who read their email, remembering a product they used last Tuesday. A question inside the product reaches the people using it, while they are using it.

1. Start from the template

In Respondo, choose Create survey in the sidebar, then the Net promoter score template. It asks two questions:

  1. 1How likely are you to recommend us to a friend or colleague? — on the 0–10 scale, labelled from “Not at all likely” to “Extremely likely”.
  2. 2What's the main reason for your score? — an open answer.

Keep the second question. The score tells you where you stand; the reason is the part you can act on. You can reword either one on the Questions tab, and mark the open question optional with Allow skipping if you would rather not require it.

2. Ask the right people, at the right moment

NPS measures a relationship, so ask people who have one with your product. On the Display tab:

  • Pages — target where regular users spend their time, such as the dashboard, rather than the sign-up flow or the pricing page. Rules can match a URL that contains, equals or starts with a path, or a regular expression.
  • How many people see it — sample a share of visitors instead of everyone. Each visitor's place in the sample is rolled once and remembered, so it does not change between visits.
  • Delay — wait a few seconds after the page loads, so the question does not interrupt someone mid-click.
  • Ask each person only once — on by default. Once someone answers, the survey does not show again in that browser.

On the Appearance tab, match the widget to your product: the accent and background colours, the corners, the width, and where on the page it sits. A survey that looks like part of the product gets answered like part of the product.

3. Put it live

Choose Embed on the survey. It shows the snippet with your publishable key and the survey's id already filled in. Paste it before the closing body tag — once, in a shared layout, covers every page — and choose Publish.

index.html

<script
  src="https://app.myrespondo.com/embed.js"
  data-key="rsp_pk_…"
  data-survey="…"
  defer
></script>

If you know who is signed in, pass it along, so a detractor's reason arrives with a way to follow up:

after sign-in

window.Respondo.identify({
  email: user.email,
  traits: { plan: user.plan },
});

4. Read the reasons, not just the score

Answers arrive as people give them. Overview shows your totals, and Responses lists every answer as it comes in. For the score itself and the split between the three groups, ask your assistant through the MCP server — get_results works out the NPS for you.

Then read the open answers, and read them by group:

  • Detractors tell you what is broken. A detractor who explains why is the most useful response you will get.
  • Passives tell you what is missing between good and great.
  • Promoters tell you what to protect, and the words your best customers use for what you do.

Common mistakes

  • Asking too early. Someone on their first day cannot tell you whether they would recommend you.
  • Asking everyone, all the time. Sample a share of visitors and leave Ask each person only once on.
  • Watching only the number. A score moves slowly and says nothing about why; the open answers do.
  • Changing the question. Reword it and your new scores no longer compare with your old ones.
The score tells you where you stand. The reasons tell you what to do next.

Your users already have the answer. Ask them.