Practical guide

NPS survey best practices: timing, sample size, follow-up and closing the loop

Last materially reviewed 2026-10-03

Quick answerSend the NPS question after a real experience, keep the standard 0 to 10 wording, report the response count beside the score, ask for the reason, and reply to every detractor.

The five NPS survey best practices that matter most

The NPS survey best practices that make the biggest difference for a small organization are these: send after a real experience, keep the standard question and scale, always show how many people answered, ask one follow-up for the reason, and reply to low scores. The rest of this page explains each one.

  • Send when the customer can judge, not when it suits your calendar.
  • Keep the 0 to 10 scale and the recommendation wording unchanged.
  • Report the count of responses next to every score.
  • Ask one open question: what is the main reason for your score?
  • Close the loop with detractors, and tell customers what you changed.

This page deliberately gives no target score or typical response rate. Those figures vary by industry, audience and channel, and we did not find a primary source we could check today.

Timing: when to send an NPS survey

There are two kinds of NPS survey, and they need different timing. A transactional survey follows a specific event, such as a finished job, a delivery or an appointment. Send it soon after, while the memory is fresh, but only once the customer has what they paid for. A relationship survey asks about the business overall and goes out at a regular interval to people who have been customers for a while.

NPS survey timing by situation
SituationSend whenAvoid
Completed service jobOne or two days after completionBefore the work has been used
Product deliveryA few days after it arrivesAt checkout or on dispatch
Ongoing subscriptionAt a fixed interval once the customer is establishedDuring the first days of a trial
After a complaintAfter the complaint is resolvedWhile the case is still open

Do not ask the same person too often. Pick a minimum gap between surveys for each customer and keep to it. More on this in survey timing.

Sample size: why a small NPS swings so much

NPS is a difference between two percentages, so it moves sharply when few people answer. With 20 responses, each person is worth 5 points. If one customer who would have given a 9 gives a 6 instead, the score falls by 10 points. With 200 responses the same change moves the score by 1 point.

Here is that sum worked through. Twenty responses with 10 promoters and 4 detractors gives 50 percent minus 20 percent, an NPS of 30. Change one promoter to a detractor and it is 45 percent minus 25 percent, an NPS of 20.

  • Write the count beside the score every time: "NPS 30 from 20 responses".
  • With small numbers, report the counts themselves: 10 promoters, 6 passives, 4 detractors.
  • Do not compare two periods unless both have a similar and reasonable number of responses.
  • Remember who did not answer. People who reply may not be typical of all your customers.

Follow-up: ask why, then sort the comments

The score tells you where you stand. The comment tells you what to do. Ask one open follow-up straight after the rating, and make it optional so the rating itself is not lost if someone skips it.

Then read and group the comments. A simple method is to give each comment one or two topic labels, such as price, speed, staff or quality, and count how often each appears among detractors and among promoters. The steps are in comment coding. Some tools do a first pass for you. Yay! Forms lists automatic keyword extraction and sentiment detection on every plan. Treat automated labels as a starting point and check a sample by hand.

Closing the loop with detractors, passives and promoters

  1. Set an alert for scores of 0 to 6. Someone should see a detractor answer the day it arrives.
  2. Reply personally. Thank the customer, say what you understood and say what happens next. Do not argue with the score.
  3. Record the outcome. Note what was wrong and whether it was fixed, so repeated problems show up.
  4. Ask passives what was missing. A 7 or 8 means something held them back from a 9.
  5. Thank promoters. A referral or review request is reasonable here.
  6. Tell everyone what changed. A short note saying what you fixed gives people a reason to answer next time.

Collecting scores and never replying wastes the survey. The full process is in closing the feedback loop.

When a paid NPS tool is worth the money

A free form and a spreadsheet handle the question and the sum. A paid tool is worth it when you need the alert and the sorting done for you. As an example, Yay! Forms calculates NPS automatically on every plan. Its pricing page includes Slack, WhatsApp and email notifications in the list of features in all plans but names them only on the Plus and Professional cards. Plus was $39 a month billed monthly or $29 billed yearly when read on 3 October 2026. If alerts on a low score are your main need, confirm in the trial that the plan you are buying includes them. Our comparison of NPS survey tools lists the alternatives.

Sources used for this page

The facts above come from the pages below, read on 3 October 2026. We have not used these products hands-on. Prices and terms change, so confirm them on the vendor's site before you buy.

  1. Measuring Your Net Promoter Score | Bain & Company — Independent · netpromotersystem.com · checked 2026-10-03
  2. Yay! Forms Pricing — Vendor pricing page · yayforms.com · checked 2026-10-03
  3. NPS Surveys with AI Analysis | Yay! Forms — Vendor product page · yayforms.com · checked 2026-10-03