See which complaints cost you points. And how many.
Upload your NPS or CSAT export. See each theme's effect on your score, how many comments it rests on, how sure the number is, and the comments behind it.
First analysis free. No card. No sales call.
What moves your rating
Sample data, clothing reviewsWorks with exports from SurveyMonkey, Typeform, Qualtrics and any CSV
Same data, same numbers
Every number links to its comments
First analysis free, then monthly
Frequent is not expensive
Tagging tools count. We measure. The theme your customers mention most is often not the one that costs you most.
Complaints customers mention most
- 1Fit or size32 %
- 2Doesn't look flattering15 %
- 3Fabric10 %
What actually costs you points
- 1Doesn't look flatteringup 1−0.14
- 2Fit or sizedown 1−0.09
- 3Fabric−0.06
Share of reviews on the left. On the right, how many stars each theme costs the average rating: the effect per mention times how often it comes up.
Every number, with its evidence
Each theme shows its effect per mention with a likely range, what it costs your average, how many comments it rests on, and whether that is sure enough to act on. Open a row to read the comments.
What moves the rating
4,000 women's clothing reviews, 1 to 5 starsData: Women's E-Commerce Clothing Reviews, public domain (CC0). An effect is an association, not proof of cause. We say so, because the others don't.
Why not just ask ChatGPT or Claude?
We tried. Given the same reviews twice and asked what drives the rating, a chat model finds sensible themes. The numbers are the problem: they change from run to run, and they mix each theme with everything else the unhappy customers wrote.
Use a chat model for ideas. Use Leverboard for numbers you can report.
| What you need | Chat prompt | Leverboard |
|---|---|---|
| Same numbers next month | No, they change between runs | Yes, same data gives the same result |
| Effect of each theme on its own | Raw differences, themes mixed | Each theme held apart from the others |
| How sure the number is | Not stated | Count, likely range, low data marked |
| Click to the comments | No labels per comment | Every number opens its comments |
How it works
No setup project and no taxonomy to build first.
- Upload your export
A CSV or Excel file from your survey tool, with a score and a comment per response. We find the columns.
- We read every comment
Each comment is sorted into themes for your industry. Themes your customers raise that are not on the list show up as suggestions.
- You get the ranked report
Effect, comments and how sure, by email within one working day. Each month, with the change since last month.
Works from a few hundred responses
With few mentions the range is wide, and we show it. Themes with too little data are marked, not hidden, so you know which numbers to wait for.
Pricing
Public prices. Monthly. Cancel any month.
Your first analysis is free
Up to about 1,000 comments. The full report with effects, ranges and comments, by email within one working day. No card.
Try it on your dataPro
- One survey program
- A report every month, with the change since last month
- Alerts when a theme grows
- Themes edited to your wording
- Full export: report and the theme of every comment
- Coming: SurveyMonkey and Typeform connection, MCP server
Team
Everything in Pro, plus
- Several surveys or locations
- Segments by plan, region or location
- Several users
No sales call. Your price stays your price. Polar Software Inc. sells the plans as merchant of record and handles payment, invoices and tax.
Questions
Do I need to talk to anyone?
No. Upload your export, and the report arrives by email. There is no demo and no sales call.
Which survey tools does it work with?
Any tool that exports a CSV or Excel file with a score and a comment per response: SurveyMonkey, Typeform, Qualtrics, Delighted, Google Forms and others.
What does the effect of a theme mean?
It is how much lower or higher the score is, on average, when a comment raises that theme, with the other themes in the same comment held apart. The method page explains what each number means.
Does a theme cause the drop?
Not necessarily. An effect is an association in your data, not proof of cause. It tells you where to look first, and the comments behind it tell you what is going on.
How many responses do I need?
A few hundred responses with comments give useful results for the common themes. With fewer, the ranges get wide, and themes with too few mentions are marked as low data.
Which languages are supported?
English comments. Other languages are not tested yet.
Is my data used to train AI?
No. Your comments are processed only to produce your report and are deleted when you ask, or after 90 days.
Why is it cheaper than CX analytics suites?
Because it needs no sales team, no onboarding project and no consultants. You upload, you read the report.
Find out what your comments cost you
First analysis free. Upload takes a minute. No card.