What is MaxDiff, and why is it such an important tool in capturing customer preferences? That’s what we answer in this, our first of a series of posts on MaxDiff.

In any given business scenario, providing customers every possible feature or service is very costly. Marketers need to prioritize their offerings on the basis of customer preference. Effectively capturing this preference is crucial and can be done in multiple ways.

The simplest approach is the scale method, which has customers rank or rate a list of features. But this comes with a significant downside. If customers are asked to share their preferences for a list of features without any trade-off, they are likely to give a high rating to everything on the list. Obviously, this does not help much. And in most cases, it can be almost pointless. For example, hospitality or travel feature lists are often quite long; this tends to cause preference ranking and other traditional methodologies to fail.

Clearly, there is a need for more advanced market research techniques in these scenarios.

Finding the Best Market Research Techniques

There are a host of market research techniques used to understand customer preferences. MaxDiff and conjoint analysis are just two of them. Depending on the research need, these techniques can be employed alone (to pinpoint certain facts) or together (to draw more insights).

For example, suppose you want to see how a service should be bundled or priced with respect to a broad list of services. In this case, conjoint analysis is the best approach. To focus on market preferences from a single list—for instance, which service benefits are most important to your audience— then MaxDiff is the go-to technique.

What Is MaxDiff?

MaxDiff, also known as Best-Worst Scaling, provides a hierarchy of the relative importance of a list of features. MaxDiff offers the researcher these distinct advantages over other methodologies:

  1. Longer feature lists (10 or more) can be handled easily
  2. The order and differentiation of features is clear
  3. Preference strengths are highlighted

For example, instead of saying that Feature A is favored over Feature B, MaxDiff reveals that Feature A is twice as important as Feature B.

  1. Responses are easier for users; they only need to judge the best and worst options, so they don’t have to spend time ranking things they may not care about

MaxDiff is widely used, thanks to its simple design for respondents and marketers alike. But how does MaxDiff actually work? Let’s look at a short sample survey to find out.

MaxDiff At Work

As a part of a MaxDiff exercise, respondents are shown multiple sets of features. For each set, they are asked to select:

  • The most important (or most appealing) item, and
  • The least important (or least appealing) item

Please select the loyalty program benefit that is most important to you and the one that is least important from the following options:


From a list of 20 or 30 features, MaxDiff can determine which preference order of all the features. Here’s what this could look like:


So, we can see how MaxDiff has provided us with an excellent measure of relative importance for different features. The researcher now understands what will work with customers and what won’t.

MaxDiff is a powerful tool when a product or service needs to be included or eliminated. But there is a point where a simple Best-Worst Scale can’t help you. This occurs when the absolute preference isn’t always attainable. If a business wants to bifurcate a feature list so that they can create bundles of features in addition to the preference order, a traditional MaxDiff can’t do the job. There may also be a situation where the features on a particular screen are completely irrelevant to the respondent or all are equally important. There is no way of finding this out with MaxDiff because there is no absolute threshold.

To get around these limits, there are variants of MaxDiff now available to marketing researchers. In the course of our next few blog posts, we’ll examine the whats and whys of all these techniques. Be sure to join us!

Authored by Rajat Narang, Associate Director of Analytics at Absolutdata and Aaroshi Asija, Senior Analyst, Market Research at Absolutdata