How to Choose a Survey Rating Scale: 5, 7, 10, or 11 Points?

Key takeaways from the article

The rating scale in a survey is a methodological decision that directly impacts the quality of the collected data, its interpretation, and its business value. Before designing your next CX survey, consider the following guidelines.

• There is no single “best” rating scale—the choice of scale depends on the survey’s objective and the type of metric (NPS, CSAT, CES, relationship survey, UX).
• For NPS, use only a 0–10 scale (11 points). For satisfaction questions in CX surveys, a 5- or 7-point scale usually works best. In UX research and attitude assessments, a 5- or 7-point Likert scale works well.
• Changing the scale during a CX program makes it difficult to compare results and analyze trends—it’s better to plan the appropriate scale from the start.
• A CX platform, such as YourCX, facilitates the standardization of scales, top-box/bottom-box analysis, and the combination of numerical ratings with customer comments.

Introduction: Why the choice of rating scale in a survey matters

The rating scale in a survey is not just a minor detail of the form. It is an element that influences how customers respond, what the distribution of results looks like, whether segmentation is possible, and how accurate business decisions based on this data will be.

Imagine this scenario: an e-commerce team measures CSAT on a 1–5 scale on the website, while the call center uses a 0–10 scale after a phone call. Comparing results across channels becomes unreliable. Management sees different numbers, doesn’t know whether phone support is better or worse, and the analyst wastes time on adjustments instead of taking action.

The wrong scale leads to misleading averages, false trends, distorted results, and reporting issues. This article answers the question: which rating scale to choose—5, 7, 10, or 11 points—depending on the survey’s objective, metrics, and channel.

What is a rating scale in a CX survey

A rating scale is an ordered set of responses that translates the respondent’s subjective experience—satisfaction, effort, loyalty—into numerical data. This enables analysis, comparison of results, and operational decision-making.

In survey research, we distinguish between different types of scales. A nominal scale has no natural order of values (e.g., selecting a product category). An ordinal scale allows values to be ordered but does not measure the differences between them. An interval scale enables the calculation of differences between values—most rating-type survey questions are based on this scale. A ratio scale has a natural zero point and allows for all mathematical operations, but it rarely appears in opinion surveys. A separate type is the dichotomous scale (yes/no), as well as the semantic scale, including the semantic differential, where the respondent evaluates a given topic on a continuum between two opposing adjectives.

In practice, CX uses numerical scales (e.g., 1–5, 1–7, 0–10), descriptive (e.g., “very bad” – “very good”), and mixed (numbers and verbal labels). Example of a 1–5 CSAT question: “How would you rate the quality of chat support?”. CES 1–7: “How easy was it to resolve your issue?” NPS 0–10: “How likely are you to recommend our company to friends?” Process rating 1–10: “How would you rate the overall online shopping experience?”

The number of points alone is not enough—the labels for the extremes, the presence of a neutral point, and consistency with other questions in the survey are equally important. Multiple-choice questions or closed-ended questions with a single answer follow different rules than rating scales.

Key criteria for choosing a rating scale in a survey

The choice of scale in a survey depends on the research objective—this is a rule worth starting with when designing any CX survey. Below you’ll find criteria to consider before selecting the appropriate scale.

• Survey objective —NPS monitoring, a quick post-purchase CSAT, or a UX study of the checkout process all require different scales.
• Metrics type —NPS, CSAT, CES, process evaluation, product evaluation—each metric has its own standards.
• Required granularity —high for strategic research, lower for transactional surveys.
• Survey length and channel —the burden on the respondent varies across online surveys on mobile, email, SMS, in-store kiosks, or IVR hotlines.
• Respondent type —B2C, B2B, seniors, users accessing the site via a mobile app—each group has a different tolerance for complexity.
• Survey frequency —continuous NPS, monthly CSAT, quarterly relationship survey.
• Comparability —should results be comparable over time, across countries, brands, or channels?
• Ease of interpretation – will operational teams and management understand the results without a lengthy explanation?
• Analytical maturity —does the organization use medians, distributions, top-box scoring, or predictive models?

There is no single universal scale. The same organization may use a 0–10 scale for NPS and a 1–5 scale for transactional CSAT questions.

The 5-point scale—when to use it

The 5-point scale is the most intuitive and quickest to understand. For this reason, it is often called the “gold standard” in research—the 5-point scale is the gold standard especially where response rate and simplicity are priorities.

Advantages: low cognitive load, well-suited for small screens and short surveys. Easy communication of results (e.g., percentage of 4–5 ratings). Widespread acceptance among customers with varying levels of digital experience. Simple surveys are more likely to yield reliable responses. Works well in transactional research: post-purchase CSAT, post-customer service interaction, post-delivery.

Limitations: less detail than a 7-point scale. Greater concentration of results in the highest categories (4–5), making it harder to detect subtle changes. A scale that is too short limits the expression of nuanced opinions. Potential for blurring differences between segments (e.g., new vs. regular customers).

Examples: “How would you rate the delivery time of your order?” (1–5: very poor – very good). “How satisfied are you with how your issue was resolved by the hotline?” (1–5). “How would you rate the clarity of information on the product page?” (1–5).

7-point scale – a compromise between simplicity and precision

The 7-point scale offers greater resolution than the 5-point scale, while remaining acceptably simple. The 7-point scale captures more subtle differences in opinions, which is essential for analyzing segments and trends.

Applications: UX research (evaluation of forms, shopping carts, customer dashboards), brand attitude surveys, measuring Customer Effort Score (CES) on a 1–7 scale in SaaS services, periodic relationship surveys in B2B, as well as in the context of usability testing.

Advantages: better differentiation of responses—the ability to capture nuances (e.g., responses such as “somewhat satisfied” vs. “very satisfied”). Greater precision in segmentation (loyalty scoring, churn risk groups). A study by Altuna & Arslan showed that the dimensional structure of the scale (factor analysis) was better suited to a 7-point scale.

Limitations: requires careful description of the extreme values and the midpoint. In very short transactional surveys, it may be perceived as “too detailed.” In channels with very short attention spans (SMS, push notifications), it may lower the response rate.

Examples: CES: “To what extent do you agree with the statement: ‘Resolving this issue required little effort on my part’?” (1–7: strongly disagree – strongly agree). UX: “How would you rate the ease of finding information about returns on the website?” (1–7).

The 10-point scale—when it adds value and when it complicates analysis

The 10-point scale (1–10) is a popular choice because it seems “precise” and familiar. The 1–10 scale is often associated with school grading, which gives respondents a sense of intuitiveness, but in CX it creates interpretation issues.

Advantages: a high number of possible ratings, attractive when presenting averages (e.g., 8.4/10). It can be useful in one-off studies, such as when evaluating a new product.

Limitations: lack of a natural midpoint (the point “between” 5 and 6 is counterintuitive). Difficult to interpret—does the difference between a 7 and an 8 in a respondent’s rating represent a real change in experience? High risk of differing response styles across cultures. Too many options can lower data quality. A scale that is too long can be difficult to interpret. A study by Dawes (2008) showed that means on a 10-point scale are statistically lower than those rescaled from 5- and 7-point scales—the difference was approximately 0.3 points.

A 10-point scale may make sense in internal surveys with established interpretation standards or as an auxiliary scale in relational research. Unless there are strong historical reasons, it is better to consider a 5- or 7-point scale.

The 11-point scale 0–10—why it is the standard in NPS

The 0–10 scale is the standard in NPS surveys. The classic Net Promoter Score uses 11 possible responses and serves as the basis for measuring customer loyalty worldwide.

Customer segmentation in NPS: 0–6 are detractors (critics), 7–8 are passives, 9–10 are promoters. Formula: NPS = % of promoters – % of detractors.

The difference between NPS and a standard 0–10 rating: NPS is a specific question about the likelihood of recommendation, a specific segmentation, and a specific calculation method. A simple average of 0–10 responses is not NPS and does not provide comparable metrics.

Advantages: industry standard—easy comparison of results with other companies and markets. Clear segmentation of customers into groups useful for retention campaigns. Works well with the open-ended question “What is the main reason for this rating?”

Limitations: NPS does not directly measure satisfaction with a single interaction. The NPS number alone does not explain “what to improve”—it requires additional questions and analysis of comments. With a small sample size, it can be highly variable. A study by Menold & Toepoel (2024) showed that on mobile devices, the 11-point scale is fully visible in only about 59% of cases, compared to 99% for the 5-point scale.

The Likert scale—when to measure agreement, attitude, or opinion

The Likert scale was developed in 1932 by Rensis Likert, an American psychologist—the name comes from its creator, Rensis Likert. It is one of the most popular tools for measuring attitudes and opinions in social research. The Likert scale is the most commonly used scale in social research, including marketing research, opinion polls, and for thesis projects.

The respondent indicates the degree to which they agree with a given statement—e.g., “The purchasing process was simple for me.” A typical Likert scale has 5 or 7 response options: ranging from “strongly disagree” through “no opinion / somewhat agree” (I somewhat agree) to “strongly agree” (strongly agree, agree strongly). Variants of the Likert scale include 5- and 7-point versions, and the Likert scale may have a neutral option in the middle of the scale (e.g., “I don’t know, I somewhat agree” → “I somewhat agree”).

Applications: brand attitude research (trust, recommendation), UX research (ease of use, message comprehension), assessment of customer attitudes and beliefs, and customer experience surveys.

The Likert scale allows for statistical analysis of survey data and is used to measure attitudes and opinions. The Likert scale is a popular tool in social research, but not every numerical scale is a Likert scale—the Likert scale pertains to agreement with statements, not arbitrary ratings. It is worth using when combining multiple questions into an index and analyzing it over time. The respondent should clearly understand what each level means.

Even or odd scales—is a neutral point necessary?

The issue of a neutral point is a common dilemma when designing a good survey.

Odd-numbered scales (5, 7) offer the middle option as a “neutral” response. Neutral points are desirable in odd-numbered scales. Even-numbered scales (4, 6) force a directional choice—somewhat positive or somewhat negative.

When a neutral point is needed: when asking about an attitude and the respondent may genuinely “have no opinion”; in relational research (e.g., early stages of brand engagement); for sensitive questions, where forcing an opinion increases the risk of random responses. The absence of a neutral option may force false choices.

When it can be omitted: regarding a specific experience just experienced (e.g., immediately after interacting with staff); when the goal is an operational decision (e.g., identifying dissatisfied customers); in very short surveys.

There is a risk of overusing neutral responses—this is the phenomenon of the central tendency bias. It can be counteracted with clear questions, avoiding overly general issues, and adding a “not applicable” option instead of the midpoint of the scale.

Scale labels: why descriptions are just as important as numbers

A well-designed scale should be unambiguous and symmetrical—this is crucial. Numbers alone are ambiguous, which is why labels are the foundation of consistent interpretation.

Three approaches to labels:

• Descriptions only for extreme values —e.g., 1 = very dissatisfied, 5 = very satisfied; intermediate numbers without words.
• Describing each value —e.g., “strongly disagree,” “somewhat disagree,” “neither agree nor disagree,” etc.—provides a numerical value linked to an unambiguous description.
• Describing response groups —e.g., 1–2: negative, 4–5: positive; used more in reporting than in the survey itself.

Labels tailored to metrics: CES 1–7: 1 = very difficult, 7 = very easy. CSAT 1–5: 1 = very dissatisfied, 5 = very satisfied. NPS 0–10: 0 = definitely will not recommend, 10 = definitely will recommend.

Maintain consistent labels across your entire CX program. Avoid unbalanced scales (with a predominance of positive descriptions), as they inflate results and make them harder to interpret.

How the choice of scale affects survey result analysis

Different scales (5, 7, 10, 11) generate different distributions of responses across categories, which affects: the mean and median, the proportion of positive ratings (top-box, top-2-box), the proportion of negative ratings (bottom-box), the percentage of promoters and detractors (in NPS), and sensitivity to minor changes in the experience.

When you calculate the average of responses on a 5-point scale and compare it to the average of a 10-point scale, you may draw misleading conclusions. Changing the scale during a CX program (e.g., from 5 to 7 points) requires parallel measurement or analytical adjustments. Business reports should not rely solely on the average—it is worth showing the distribution of ratings, top-box scores, and segmentation.

More detailed scales (7, 10) provide more data for statistical analysis (e.g., correlations), but require greater methodological discipline and clear rules for interpretation.

Can results from different scales be compared?

Directly comparing results from different scales (e.g., an average of 4.2/5 and 8.3/10) is risky. 4/5 (80%) is not the same as 8/10 (80%), because respondents use the 5-point and 10-point scales differently, and their responses generate a different distribution and “weight” of extreme ratings.

You can use percentage conversions, but always with caution and a clear description of the assumptions. Recommendation: maintain a single scale for the same metric and process (e.g., post-purchase CSAT always 1–5). If you need to change the scale, mark the change point in reports and conduct parallel measurements across several survey waves.

Which scale to choose for NPS, CSAT, and CES

5- or 7-point scales are most commonly recommended for the majority of CX metrics. Specific recommendations are provided below.

• NPS: the classic 0–10 scale (11-point). Only this ensures comparability with the market and the literature.
• CSAT: mainly a 1–5 scale for transactional surveys. A 1–7 scale as an option for more analytical research (e.g., relational CSAT in B2B).
• CES: often 1–5 or 1–7. A 7-point scale works well when we want to differentiate effort more precisely in complex online and SaaS processes.
• Online process evaluation (registration, checkout): 5 points for simplicity, 7 points when the process is critical.
• Overall experience rating: 5 points (easy to report) or 10 (with a clearly defined interpretation).
• UX/attitude surveys: typically a 5- or 7-point Likert scale, with a set of questions forming an index.

Practical comparison table of rating scales

The table below summarizes the key differences between the scales used in CX surveys.

The table is for reference only—in practice, consistency within a given CX program is more important than an “ideal” scale.

Which scale should you choose depending on the research objective?

The table below shows recommended scales for specific business cases.

It’s worth establishing an internal rating scale standard for the most important CX metrics—and applying it consistently.

Common mistakes when choosing a rating scale

Below is a list of common issues we observe in Voice of Customer programs:

1. Frequently changing the scale without a plan —the lack of parallel measurement makes it impossible to compare results over time.
2. One scale for everything —e.g., 0–10 for every question, which is excessive for simple transactional questions.
3. Lack of descriptions for extreme values —this leads to different interpretations of “1” and “5” by respondents.
4. Confusing NPS with any 0–10 rating —NPS requires a specific question about recommendation and specific segmentation.
5. Comparing results from different scales on a 1:1 basis—e.g., 4.5/5 vs. 9/10 without methodological adjustment.
6. Analyzing only the average —without distribution, top-box, and bottom-box data, the average can mask polarization of opinions.
7. Scale too complicated on mobile – drop in response rate because options require scrolling.
8. Lack of an open-ended question —a numerical rating without comments does not explain the reasons behind extreme ratings.
9. Lack of consistency across channels and countries —a different CSAT format in the app and on the website skews the results.
10. Unbalanced scale —more individual positive categories than negative ones inflates the results.

How a CX platform helps standardize scales and interpret results

Standardizing scales in a large organization is a logistical challenge. A CX platform, such as YourCX, can help in several ways:

• Defining and maintaining standard response scales for NPS, CSAT, CES, and other metrics.
• Ensuring scale consistency across online surveys on various channels (web, app, email, SMS).
• Supporting trend analysis over time, even with planned scale changes.
• Segmenting results by channel, product, campaign, or customer profile.
• Combining numerical results with open-ended comments (e.g., text analysis for promoters and detractors).
• Automatically calculating top-box, bottom-box, and NPS scores and displaying them on dashboards.

Standardizing scales on the platform reduces the time from data collection to operational action and facilitates comparing results across teams.

Checklist: How to Choose a Rating Scale for a CX Survey

Before launching the survey, go through the following questions:

• What exactly are we measuring (loyalty, satisfaction, effort, opinion about a process element)?
• Is it NPS, CSAT, CES, a UX survey, or a relationship survey?
• Is simplicity for the respondent a priority, or greater precision for analytics?
• Through which channel will the survey be conducted (mobile, desktop, call center, store)?
• Should the results be comparable to previous surveys or across countries?
• Will respondents understand the scale well (e.g., 0–10 vs. 1–5)?
• Do we have clear labels for the extreme values and, if applicable, the midpoint?
• Do we know how we will interpret the results (e.g., what we will consider a “good” CSAT)?
• Do we plan to include an open-ended question asking for the reasons behind the rating?
• Is the scale consistent with other surveys in our CX/VoC program?

FAQ - Practical Questions About Rating Scales in CX Surveys

Below are answers to questions that frequently arise when designing research programs.

How can we reconcile different rating scales inherited from previous studies?

First, select the target rating scale standard (e.g., 0–10 for NPS, 1–5 for CSAT) and gradually migrate all surveys to this standard. Implement a transition period with parallel measurement on both the old and new scales at key points in the customer journey. Clearly mark the point of scale change in reports to avoid overinterpreting spikes in results.

What should you do if respondents report that the scale is unclear?

Simplify the scale (e.g., switch from 10 to 5–7 points) and clarify the labels for the extreme values and the midpoint. Test the survey on a small sample before rolling it out on a large scale. An unclear scale reduces data quality more than a smaller number of points.

Should scales be the same across all channels?

For the same metric (e.g., post-contact CSAT), it’s best to use the same scale across all channels to make it easier to compare results. Minor differences in presentation format (e.g., stars on mobile, numbers in email) are acceptable, provided the number of levels and labels remain consistent. Centralized management of survey templates, e.g., in a CX platform, reduces the risk of inconsistencies.

How often can the scale be changed in the CX program?

The scale should remain as consistent as possible. Changes should only be made when there is a clear reason to do so (e.g., integration of research programs, internationalization). With every change, you must: clearly indicate this in reports, plan a comparison period (parallel measurement), and update KPI definitions for operational teams.

Are graphical scales (stars, smiley faces) just as good as numbers?

Graphical scales can improve the survey’s UX (especially on mobile), but behind the scenes, they should be mapped to a consistent numerical scale (e.g., 1–5). More important than the visual form is the consistency of numerical values and their interpretation across channels. In documentation and internal reports, always use an explicit numerical scale, regardless of whether the client sees stars or numbers.

Summary: Which rating scale should you choose for your survey?

The best rating scale for a survey depends on the survey’s objective, not on a universal rule. A 5-point scale offers simplicity and a high response rate. A 7-point scale provides greater precision and works well in analytical research. A 10-point scale requires careful interpretation and clear thresholds. The 0–10 scale (11 points) is the NPS standard—use it exclusively with the classic recommendation question.

Final tips:

• Consistency over time and across channels is more important than the “perfect” score.
• Clear labels, avoiding unbalanced scales, and combining numerical ratings with customer comments are the foundations of a good survey.
• The Likert scale most often has 5 or 7 points and works well in attitude, UX, and opinion surveys.

Start with the question: what business decision is the survey meant to support? Only then should you select the appropriate scale that will “feed” that decision with reliable data.