Consistency in the scale of customer survey questions is important, not only across your organization, but throughout a survey. Consistency in scales by feedback type, e.g. relational or transactional, across your organization enables comparable data analytics. Consistency in scale throughout a survey contributes to the overall ease of survey taking for the respondent, reducing the potential for survey frustration.

Greater granularity in analysis is valuable, particularly for predictive models. For that reason, we recommend standardizing on the 11-point (0 to 10) scale used in the Likelihood to Recommend question that forms the basis of the Net Promoter Score (NPS). Continuing to use the 11-point scale when you measure driver/attribute satisfaction in NPS relationship feedback makes analysis more reliable. That being said, the mode used for data collection, e.g. mobile app, may dictate the need for a more concise scale, either 7-point or 5-point for measuring driver/attribute satisfaction in a relationship survey.

When it comes to transactional or event feedback, in which the primary metric is likely to be satisfaction, simplicity is key. Satisfaction should be on either a 7-point (1-7 with 1 on the left) or a 5-point scale (1-5 with 1 on the left)/ 5-star rating. Having a mid-point to anchor the scale is important, thus we discourage 6-point and 4-point scales. A 3-point scale is too limited in range and hinders meaningful analytics. The trend to 5-star ratings via apps for consumer feedback is expected to continue. The use of smiley faces for customer experience ratings is problematic due to the need to have at least 5 options and interpretation of the 5 faces may vary widely by respondents.

Regardless of scale range, only the endpoints of scales should be labeled, providing clarity on high and low ends of the scale. Labeling of scale points beyond the end points introduces ambiguity and poses challenges in interpretation by the respondent. Another challenge with multiple labels is comparable translation. In general, numerical differences are easier for the respondent to understand. We discourage use of a “neutral” label in the middle of scales, and it should be noted that it was not used in the original research that established the value of Net Promoter Score.

The “not applicable” or NA option may be added as appropriate to Satisfaction or Agree/Disagree scales to give the respondent a way to indicate they are not in a position to give a rating for a specific experience or attribute.

Foundational practices related to NPS metrics look primarily at data you get when seek direct feedback from customers. It’s limited by response rates – a problem that only gets worse over time – and response bias. Taking metrics to the next level means transforming robust, always-on streams of operational data into calibrated, AI-created scores, so that you have a personalized Net Promoter Score for every customer based on their ongoing interactions with your company. To get there, download our guide, take our comprehensive training course, or read about the future of NPS.