Updating CSAT rating scale from 11 pt to 5 pt -- Do or not do / How to handle | XM Community
Skip to main content
I have been running a customer survey program for 3.5 years (read, lots of historical trending data!) where we have done everything on an 11 pt scale with 3 points labeled, to correspond with our all important NPS survey 🙂 (We have reported average score out of 10 for non NPS questions.) Now I am interested in updating to a 5 pt labeled scale for all transactional surveys to be easier on the customer taking our surveys, + strive for more valid results. Has anyone done this?



I am very interested in any of the following:

1. Info on how it went -- did you have any unexpected issues making the update; would you do it again?

2. How did you handle reporting out the results given it was out of line with the historical reporting? top box and switch to reporting % of satisfied customers? Did you attempt to normalize results for any period or just do a hard switch and have an asterisked note?

3. If anyone has lots of wisdom on this and would be willing to have a 15 minute phone chat or email back and forth please let me know, with thanks!!
I did this exact thing five months ago- I even posted the same question but no one had any insights, haha.



I had far less trouble than I thought. My main issue was convincing the execs what reasonable new targets would be. I re-calculated all my old data to assume a "high baseline" of 8-10 = satisfied, and "low baseline" that 7-10 = satisfied. Also- my new scale contains "very" and "somewhat" satisfied, both of which counts towards the end metric of "satisfied customers".



I have 8 transnational surveys we were watching- there are only 3 months of data here to work with on the new scale, but only 1 survey had a reporting period outside of that baseline range. So that was really great!



I have a few annual surveys that won't roll around until September. So I have no idea if those will react differently or not.



But overall I would do it again. The 10 point scale was a legacy item left to me that I've never really liked. Our surveys are so short that the type of regression modeling that makes it helpful is not really applicable to our work. The Likert scale fits on mobile devices better. I have too little time to go off of, but it looks like our completion rates got a little boost. And I like to think it's just less burdensome on our customers.



PM me if you want to chat any more. Happy to help if I can.

Leave a Reply


Terms & Conditions