When should I send the survey?

If it is transactional, as soon as possible: sends within 2 hours of the interaction get around 32% more complete responses. If it is relational, quarterly. More often tires people and lowers the rate without giving you new information.

Can I change the scale later?

You can, but you lose all the historical comparability you have accumulated. Pick a scale at the start and never change it: migrating from 1-5 to 0-10 erases the previous months of work.

06 · CX

Design an NPS survey you can actually read

An NPS of 42 calculated over 38 responses has a margin of about ±20 points: the true value sits somewhere between 22 and 62, and next month it can "rise to 55" without anything having happened. That number gets presented in meetings all the time, bare, and used to make decisions. The problem is not NPS: it is presenting it without an interval.

No signup · no email · runs in your browser

Your send

How many people you will survey, and when.

Sends per monthcontacts

Survey type

Send timing

Actual responsesoptional

Your result

How responses split across the 0-10 scale.

Promoters9-10

Passives7-8

Detractors0-6

General trend only

Around 73 responses a month

Enough to talk about a general trend and to read the open comments, where the actionable material lives. Not enough to compare branches, staff or segments. It yields between 50 and 80 responses from 200 sends. The chosen timing places the estimate in the upper part of that range — not an invented multiplier, just where to land inside the published range.

Expected responses

NPS margin at that n

±17.0

Possible segments

✓

Read a global NPS and present it with its margin

Compare the result against last month

Split into two segments (branch, product line)

Split into four or more segments

Analyse performance per staff member

Your NPS with its margin

The number alone, without an interval, does not tell you whether you improved or it was chance.

55%

30%

15%

Promoters · 55

Passives · 30

Detractors · 15

NPS and 95% intervalNPS +40 · ±14.5 · 26 → 54 · n=100

−1000+100

Check how you wrote it

Heuristics for the most common wording defects. It does not replace reading it out loud.

✓ No defects that this check detects. Now read it out loud: if it sounds odd, it is.

Why it exists

What you can claim with the volume you have

If the plan is 200 sends a month, between 50 and 80 responses will arrive. That is enough to talk about general trends and to read the open comments, which is where the actionable material lives. It is not enough to compare branches or staff.

The tool marks five claims as possible or impossible based on your n: reading a global NPS, comparing it against last month, splitting it into two segments, into four, or analysing by staff member. This is the part that prevents the awkward month-three conversation.

The interval uses the correct NPS formula, which is not that of a simple proportion: Var = [pP + pD − (pP − pD)²] / n. With 55 promoters, 30 passives and 15 detractors it gives NPS +40 with an interval from 26 to 54.

Questions

What people usually ask

Below 30 you do not have an NPS: you have comments, which are valuable and should be read one by one. Between 30 and 100 you can talk about general trends. Above 100 you can compare month over month, and each additional segment needs about 100 responses of its own.

Other tools

General assessment

Did you get a number you did not expect?

Send it over and I will tell you what I would do with that case in the first 30 days, with nothing to hire. If the honest answer is that you do not need an agency yet, I will tell you that too.

Let’s talk See the service