WooCommerce Post-Purchase Survey: 7 Questions, When to Ask and What to Do Next Rating scale
Glossary 
Mike Taylor Updated: Nov 11, 2026 Reading time ≈ 5 min
- Topic
- Questions and scales
- Template
- Rating scale survey
Contents
A rating scale is a survey question that asks people to score something on a set of ordered points, such as 1 to 5, five stars or 0 to 10. It turns opinions about quality, satisfaction, importance or likelihood into numbers you can compare across products, teams and time periods.
This page covers what a rating scale is, its types and how to design one. For ready label sets for 1–5, 1–7, 1–10 and 0–10 scales, see Rating scale examples.
Example of a rating scale
"How would you rate the quality of our support?"
1 Very poor, 2 Poor, 3 Fair, 4 Good, 5 Very good
Every rating scale has three parts: the thing being rated, the number of points and the labels that give each point its meaning. This example is balanced, with two negative labels, a neutral middle and two positive labels. The familiar "Poor, Fair, Good, Very good, Excellent" set leans positive, because only one of its labels is clearly negative.
Types of rating scales
| Type | What respondents see | Typical use |
|---|---|---|
| Numeric | Numbered points, labeled at the ends or on every point | "From 1 to 10, how would you rate the course?" |
| Verbal (descriptive) | Words only, such as Very poor to Very good | When every point needs a clear, shared meaning |
| Star and emoji | A row of stars, icons or faces | Quick feedback on products, orders and events, often on a phone |
| Slider | A handle dragged between two ends | Fine-grained intensity, as in the visual analog scale |
| Likert-type | Agreement with a statement, from Strongly disagree to Strongly agree | Attitudes and opinions (see Likert scale) |
| NPS 0–10 | 11 points from "Not at all likely" to "Extremely likely" | The recommendation question of the Net Promoter Score |
Rating scales are also unipolar or bipolar. A unipolar scale measures how much of one quality there is, from "Not at all important" to "Extremely important". A bipolar scale runs between two opposites with a neutral middle, from "Very dissatisfied" to "Very satisfied".
Rating scale vs Likert scale
A Likert item asks how much people agree with a statement ("Support solved my problem": Strongly disagree to Strongly agree). A rating scale is the broader term. The respondent scores an object directly, as in "Rate our support from 1 to 5". Both produce ordinal data and are analyzed in similar ways.
How many points to use
- 5 points are the most common choice. Every point can be labeled, the row fits a phone screen and people answer quickly.
- 7 points add a little detail and suit bipolar questions about agreement or ease. The Single Ease Question uses 7 points.
- 10 or 11 points are familiar from "out of ten" ratings and the NPS question. Usually only the ends are labeled, and neighboring numbers are hard to tell apart.
- Odd or even. An odd number gives a neutral middle. An even number makes people lean one way. Put "Don't know" or "Not applicable" outside the scale as a separate option.
More points add detail only when people can tell neighboring points apart. Consistency matters more than the exact number. If you compare results with an earlier wave or with a standard metric such as NPS or CSAT, keep the same points and labels.
How to analyze rating scale answers
Rating scale answers are ordinal. The points are in order, but the step from 1 to 2 may not feel the same as the step from 4 to 5. Report more than one number:
- Distribution: how many people chose each point. A mean of 3 can come from everyone choosing 3 or from half choosing 1 and half choosing 5.
- Median and mean: the median is the safer summary for ordinal data. The mean is common and handy for comparing groups when the distribution is shown next to it.
- Top-2 box: the share of the two highest points, such as 4 and 5 on a 1–5 scale. See top-box score.
The rating average calculator and the top-box calculator work these out from your answer counts.
Common mistakes
- Unclear poles. "1 = minimum, 10 = maximum" leaves people guessing. Name what is measured: "1 = Not at all useful, 10 = Extremely useful".
- Mixed directions. If 5 is the best answer in one question and the worst in the next, respondents and analysts both make mistakes.
- Unbalanced labels without a reason. More positive than negative options pull answers upward.
- Two things in one question. "Rate the speed and friendliness of support" can't get one honest score. See double-barreled question.
- Forcing a score. People who never used a feature need a "Not applicable" option.
- Reporting only the mean. Show the distribution or the top-2 box with it.
How to build a rating scale in SurveyNinja
The Scales and ratings section of the SurveyNinja builder has several question types for rating scales:
- Opinion Scale: a row of numbered buttons that starts at 0 or 1 and ends at any number from 3 to 10, with your own labels under the first, middle and last number. The report shows the average rating and splits answers into low, middle and high groups (Opinion Scale and Slider).
- Rating: stars, hearts or other icons. The report shows the average, the median and the standard deviation.
- Smile rating: two to five faces from sad to happy, with an optional follow-up question after the faces you choose (Star and Emoji Rating Questions).
- Slider: the respondent drags a handle to pick a whole number in the range you set.
- NPS: the 0–10 recommendation question. The report calculates the Net Promoter Score.
Opinion Scale, Rating, Smile rating and NPS work on every plan, including Free. To rate several items on one scale in a single grid, use a matrix question (Basic plan and higher). You can also start from the rating scale survey template.
Updated: Nov 11, 2026 Published: May 31, 2026