Trying to understand MaxDiff Preference Utility | XM Community
Skip to main content
Question

Trying to understand MaxDiff Preference Utility

  • December 21, 2020
  • 1 reply
  • 91 views
A
Level 1 ●
Forum|alt.badge.img+3

I'm trying to understand how MaxDiff Utility Score is calculated beyond what is described in:
Getting Started with MaxDiff ProjectsAnd in the white paper:
MaxDiff Analysis White PaperBasically, I want to under how do I interpret the Preference Utility? What is the significance of positive and negative preference utility. What is the significance of a 0 (zero) .


1 reply

K
Level 5 ●●●●●
Forum|alt.badge.img+19
  • Kate
  • Level 5 ●●●●●
  • December 22, 2020

MaxDiff is all relative only to itself.
MaxDiff works best when you have a definitive set of things you're trying to test against one another. Most common use case I've seen is message testing - do you prefer Ad Copy A, or B? The preference numbers themselves don't tell you much of anything- only that A out performs B among the audience you sent the test to. It's not like percentages where you can say "XX% of the audience prefers A"- it's truly only "Compared to everything else we tested- A wins".
For that reason, it's used best when you only have a set list of things you're testing. It's not something you want to add an "other" category to, or to be used to try and determine all factors the audience cares about. It only works under the assumption that A and B are the only things we care about. It has limited and specific use cases.

A
Terms & ConditionsAccessibility statement