I am sure you have seen your fair share of surveys. Businesses are putting high stakes into customer experience (CX) programs through a variety of reach-outs. The outcomes and insights generated from CX campaigns and surveys can provide a company with significant ROI to drive marketing decisions, operational changes, and strategies for growth. Among the surveys you've taken, you've probably seen all types of scaling questions from 3-point, to 5-point, to 7-point, to 10-point options. Although it seems practical to think "we'll just use whatever", a lot of thought goes on behind the scenes to arrive at the appropriate scale option in surveys. Or at least there should be some thought.
For the purpose of this post, we will break down the options into two: (1) small scales with fewer points, 5 or less and (2) large scales with many points, more than 5. Trying to determine which scale or combination of scales to use in your survey? Consider the following factors.
What device will the majority of respondents be using to take the survey?
Mobile devices and tablets has pushed the market research industry to smaller scales. Scaling questions with 3 to 5 points format better on a mobile device. Scales with a larger number of points often do not fit horizontally and can line break making the scale very confusing to the respondent. Many B2B surveys are taken on desktops which gives you more flexibility to add larger scales.
- Winner: Small scale if mobile is key.
Do your survey scores tend to gravitate into bundles?
If you have used 5-point scales in the last wave and the majority of your scores gravitate to the "4" or "5" score you may want to think about adding some additional scoring options to increase the variance between responses. Moving a 5-point scale to a 10-point scale will provide you with more insight. You may find that scores may remain in the "7" to "10" range but there is a larger difference in satisfaction between a "10" and a "7" than a "5" to "4".
- Winner: Large scale if respondents' bundle selections.
Do you need to track specific metrics?
If you are looking to track scores like NPS (likelihood to recommend) you may need to settle with the suggested NPS 11-point scale (0 to 10.) NPS is an industry benchmark score that subtracts the detractors (those rating you "0" to "6") from the promoters (those rating you "9" or "10".) I've seen some mobile friendly versions of NPS using a 4-point scale ("1" and "2" equal detractors, "3" equals passive, and "4" equals promoter) but you run the danger of an apples-to-apples comparison when you start changing scales for key industry accepted metrics like NPS.
- Winner: It depends on the metric(s).
Do you need to run advanced analysis?
Although not terribly advanced, correlation and regression analysis like more data points. Having stronger variances and more variety in data allows correlation to better understand relationships between data points, and helps with regression drivers. If you are looking to run additional statistical analysis on your data set, opt for scales with more than 5-points.
- Winner: Large scale to run advanced analysis.
How long is your survey?
This matters because in surveys running over 3 to 5 minutes, you need to consider respondent engagement more. Repetitive scales of 0 to 10 throughout a survey will likely result in fatigue and boredom and as a result, increase dropout rates. Switching between short and long scales is a great option here because it keeps the answer choices fresh. Change is good.
- Winner: Both, switch back and forth but don't make it confusing.