A lot of people have heard of MaxDiff but don't understand how it can help them learn about their customers. MaxDiff (also called best-worst scaling) is a method that allows us to learn about respondents’ evaluations of a list of items.

MaxDiff tells us how much respondents like each item in a list or how important they think each item is. Here is an example from the Discover platform by Sawtooth.

Example of a MaxDiff survey interface asking respondents to choose their most and least preferred countries to travel to.

Respondents go through several such screens. When analyzing the results, analysts combine data from each of those screens for each respondent as well as data across respondents. The output is ratio-level data telling us how important respondents consider or how much they like each of the items. MaxDiff output allows us to quantify how much more respondents like an item, or how much more important they consider it, compared to another.

MaxDiff has huge advantages over other methods like asking respondents to rate the importance of items on scales (whether numeric or Likert scales) and asking them to rank-order items. MaxDiff has the greatest advantages when:

  • You want to simplify the task for respondents. With MaxDiff, respondents only have to say which items are best and worst or most and least important. They don't have to think "is this item a 7 or is it an 8?" or "is this important or very important?"
  • You have a lot of items to evaluate. MaxDiff makes the task more manageable for respondents by showing them 4-5 items at a time. MaxDiff allows us to get evaluations of up to 40 items. This allows us to obtain much more detailed results than with other approaches. We can learn which specific items respondents like or consider important and which specific items they dislike or consider unimportant.
  • You want to avoid respondents telling you that they like everything or that everything is important. When we ask respondents to rate items they usually tell us that they like all of them or that all the items are important. The results aren't helpful for business or strategic decisions. MaxDiff tells you how important each item is relative to the others, allowing organizations to focus on what really matters to people.
  • You want to avoid differences in how respondents interpret scales. Respondents interpret traditional ratings scales differently, making it hard to compare different people’s responses to the same question.
  • You want to learn about segments of the population. With MaxDiff, it’s easy to discover segments with similar preferences or concerns.

MaxDiff is a great method to learn about how much people like various items or how important they consider each of them. It helps organizations learn what really matters to people, allowing them to focus resources on the items people like the most or consider most important.

Have you tried MaxDiff analysis? Contact us if you'd like to learn more.