FreeNo sign-upYour answers stay on your computer
Net Promoter Score (NPS) Analysis Tool
Load your survey answers and see your score with its margin of error, how it differs between customer groups, and which unhappy customers to follow up, with a diagnosis of what it means.
Analyze your survey answers
A CSV export from your survey tool: one row per answer, with the 0 to 10 score. A reason, customer ID, join date or plan make the results richer.
or drop the file here
Haven't surveyed yet?
Plan it here: whom to ask, how many answers you need, and the two questions, ready to paste.
See an example first
700 answers from the members of a chain of four gyms.
Your survey answers never leave your browser
The file is analyzed on your computer. To write your diagnosis, only summary figures (scores, shares and counts) are sent to our server; no reasons, customer IDs, dates or group names. How it works
Only need the score? Use the quick NPS calculator
Plan your survey
Three decisions, then two questions to paste into any survey tool. Come back with the answers when they are in; your earlier work stays saved in this browser.
1Whom to ask first
2How many answers you need
3When to send it
- For new customers, a few weeks after they join, while the start of the experience is still fresh.
- For everyone else, on a regular schedule, the same time each quarter for example, rather than straight after a good or bad event such as a bill, an outage or a support call. Event-timed surveys tilt the score one way or the other.
- Leave it open for about two weeks and send one reminder halfway through.
4The two questions
1. On a scale from 0 to 10, how likely are you to recommend [your company] to a friend or colleague?
0 = Not at all likely, 10 = Extremely likely
2. What is the main reason for your score?
Open text, optional
Make sure the export includes each customer's ID, or their email, and if you can, the date they became a customer and their plan. The analysis can then split scores by group and name the detractors to follow up.
Check the data
The tool has guessed which columns to use. Change them if a guess is wrong, and check the summary below before you go on.
Data check
See the first five rows of your file
Your score in detail
Overview
How people answered, 0 to 10
Detractors (0 to 6) Passives (7 or 8) Promoters (9 or 10)
By how long people had been customers
The range is the margin of error: the score for that group is likely to lie within it. Groups with fewer than 20 answers are too small to read.
By group
Detractors to follow up
The 10 lowest scores are shown here; the full list is in the Excel download.
Save your progress
Download your progress file to continue from this point any time, on any computer: choose it under "Coming back?" at the start of the tool. Each new file includes everything so far, including stage 1 when it is connected, so keep the latest one. It also holds the figures and the full list of detractors to follow up, ready to use in Excel. Your results stay in this browser only until you close it.
Not saved yet.
Take your results with you
A PDF management report to share with your team, without individual customers unless you choose to include them. The list of detractors to follow up, with their IDs and comments, is in the progress file above.
About the tool
Why it exists
The score is the start; who gave it is what you act on
Net Promoter Score asks customers one question, how likely they are to recommend you from 0 to 10, and subtracts the share of unhappy customers from the share of loyal ones. A single company-wide number hides the useful part: which customers are unhappy, and why. This tool shows the score with its margin of error, splits it by how long people have been customers and by any group you choose, and lists the detractors to follow up.
It works on its own, and as stage 2 of our customer churn playbook. If you have run the churn analysis in this browser, the tool compares the two: whether the customers who leave most are also the least happy. For the method behind the score, see our guide to Net Promoter Score.
Before you start
What your survey file needs
| Column | Needed? | What it adds |
|---|---|---|
| Score | Yes | The 0 to 10 answer. Text such as "9 - Very likely" works too. |
| Reason | Optional | The answer to "What is the main reason for your score?". It is what you will act on. |
| Customer ID or email | Optional | Names the detractors, so you can follow them up. |
| Date they became a customer | Optional | Splits scores by tenure, the same groups as the churn analysis. |
| Group | Optional | Plan, region or sales channel, to compare scores between groups. |
Any survey tool's export works; other columns are ignored. If your survey tool doesn't record join dates or plans, add them from your customer list before you start, matched on customer ID.
The method
How the tool works it out
The score. Answers of 9 or 10 are promoters, 7 or 8 are passives, and 0 to 6 are detractors. The score is the percentage of promoters minus the percentage of detractors, so it runs from −100 to +100.
Margin of error. Because a survey reaches only some customers, the score for your whole customer base is estimated as a range, at 95% confidence. Smaller surveys have wider ranges. The range covers only the chance of who happened to be surveyed, not bias from some kinds of customers being more likely to answer.
Customers who answered more than once. If the file has a customer ID or email column, the tool spots repeat answers. By default it keeps each customer's latest answer, so nobody counts twice; you can choose to keep every answer instead.
Comparing groups. Scores are worked out for each tenure group and each group you choose. A difference is called out only when it is bigger than chance alone could produce with that many answers, and at least 10 points.
The diagnosis. It describes what the pattern often points to. The score shows how customer sentiment is spread, not why; the detractors' reasons and the interviews in the next stage find out. Low scores can signal a higher risk of leaving, but a score alone doesn't predict who will leave.
What leaves your computer. Your file is analyzed in your browser. To write the diagnosis, the page sends our server the scores, shares and counts only, without reasons, IDs, dates or group names, and the server keeps nothing. Your latest result is saved in your browser, and you can clear it at any time.