What Matters Most to Your Customers? MaxDiff Analysis Will Tell You

It’s understandable to want to offer your customers everything — a million flavors of ice cream, a completely silent vacuum cleaner with insane suction, and all the stars in the sky as an add-on.

Sadly, we all have limitations (even billion-dollar businesses do), and sometimes we need to make choices. Which ones, though? To get a sense of which features and attributes of your product offering matter most to your target customers, one market research method stands out: MaxDiff analysis.

What is MaxDiff analysis?

Short for “Maximum Difference Scaling,” MaxDiff analysis helps you find out what people value most by asking them to pick their favorite and least favorite options from a list. This choice-based method is an efficient way to determine the relative importance survey respondents place on certain products, features, or attributes.

Test-takers are shown a set of product options — these could be seltzer flavors, shirt patterns, skincare product benefits, or the like — and asked to choose which are most and least important (or desirable) to them.

This methodology isn’t new. If you’re a market research history buff, you might be interested to know that MaxDiff got its start in 1987, around the time that we also got the Theory of Reasoned Action and the Theory of Planned Behavior.

At the most basic level, you calculate MaxDiff output for each item by subtracting the percentage of people who called it “Least Important” from the percentage who say it’s “Most Important.” This gets you a quick ranking with minimal math. That said, there’s plenty of opportunity for the avid market researcher to go in-depth with statistical analysis (we could nerd out on this too, but we won’t right now).

What is a MaxDiff survey, and how does it work?

MaxDiff’s ability to provide powerful insights is belied by its simplicity. A MaxDiff question on a survey usually has three columns: 1) a set of product features or attributes in the middle flanked by a 2) column of radio buttons for “Least Important” on the left and 3) a column of radio buttons for “Most Important” on the right. (Radio buttons allow you to select a single option only.)

There’s a science to choosing how many options to have per question. Show more than five attributes per set, and people have a hard time making a choice. Show just two, and you might wind up with a test that’s way too long.

One crucial aspect of any MaxDiff question: There’s no in-between option; there’s just MOST and LEAST. We’ll discuss why this matters later on.

What MaxDiff studies can tell you in the CPG space

MaxDiff analysis gives you the consumer insights you need to make a variety of business decisions regarding which product features and attributes you should prioritize. This is, of course, assuming you’re somewhere in the 100% of businesses that can’t afford to offer as many options as there are Bertie Bott’s Every Flavor Beans.

If you’re considering dropping a flavor, feature, color, etc. from your product offering, MaxDiff can help you pick the one that people care about the least. This makes it a vital complement to alienation testing, which involves trying to figure out whether a particular change to a popular product will upset customers enough to make them stop buying it.

MaxDiff works for messaging as well as for features. Claims testing is a method that market researchers use to identify which claims about a given product resonate most with the target customer base. A MaxDiff survey can point out the right messaging to put on the packaging and in ads.

Something you WON’T learn with MaxDiff: how passionately people feel about the things they select as most and least important. Do people HATE the earwax flavor jelly bean, or do they just like it slightly less than the strawberry, cinnamon, peanut butter, and cotton candy ones? You’ll never know, unless you ask something more in the realm of qualitative market research.

How is MaxDiff different from conjoint analysis?

Conjoint analysis and MaxDiff are related, but not quite the same. In any MaxDiff analysis example, you’re comparing feature by feature, whereas in conjoint analysis, you’re asking people to pick their favorite bundle of features from a set of options.

In conjoint analysis, you’re basically looking at a set of different product options with different features, not feature options for the same type of product. This market research method often takes a menu-based approach, where you can select different combinations of colors, flavors, prices, and more just as you would choose an appetizer, an entrée, and a dessert from a prix-fixe menu.

Unlike conjoint analysis, MaxDiff doesn’t really evaluate pricing options, but it can be combined with Van Westendorp Pricing Model questions to understand whether certain features should be offered by default, as a premium, or as an add-on (or not offered at all).

The importance of trade-offs in MaxDiff

Like we mentioned above, MaxDiff questions have no in-between option. Your test-takers get two choices: Most Important and Least Important.

Why does this matter? Well, it boils down to the very purpose of the study, which is to reveal a select number of clear standouts in each set of options. If you ask people to rate things on a scale of 1-5, some options could end up with the same rating, which isn’t very helpful.

Similarly, if you ask people to rank things in order from favorite to least favorite, they might spend too much mental energy wondering if they should put one option above another option halfway down the list rather than focusing on the two extremes. This causes ranking fatigue, which ultimately defeats the purpose.

By keeping things simple and forcing survey-takers to make a choice, a MaxDiff survey tells you what your customers value most from your products.

How to extend MaxDiff to get even more insights

Let’s get back to the idea of how passionately people feel about each option — in an absolute sense, not just relative to the other options. A regular MaxDiff question won’t answer this for you, but you can add a follow-up question to determine which options fall into categories like “can’t live without,” “nice to have,” “meh,” and “this flavor makes me want to barf.”

This is known as Anchored MaxDiff. It can give you a sense of the overall enthusiasm about each option, and it can also help you weed out people who don’t really care about the product category as a whole. For instance, if you’re trying to figure out which TV shows to offer as part of a bundle, you probably wouldn’t want to collect a bunch of data from people who never watch TV.  

You can pair MaxDiff analysis with sensory testing to gain deeper insights into consumer preferences and product attributes. The detailed feedback you get from sensory testing, such as taste, texture, aroma, and appearance, can help you build flavor profiles for your target customer base, and then a MaxDiff survey can help determine how much more important flavor is relative to factors like texture or packaging.

What you need in order to do MaxDiff analysis well

For any type of market research, you want to have a good, representative sample size. This is no different with a MaxDiff study, in which a sample size of at least 200 respondents is considered ideal. If you’re segmenting your study into subgroups, it’s a good idea to have 200 respondents per subgroup. You might also need a larger sample size if you’re testing an extra-large number of feature/attribute options.

A solid understanding of how to design the surveys and interpret results is also vital. If shown a MaxDiff example alongside a conjoint analysis example, would you know which was which? Do you know how to choose attributes and make your questions clear, clear, distinct, and measurable? (Clarity is essential in writing good market research questions.)

Highlight can help with these things. Highlight’s ability to put together a representative audience that can also be segmented into highly niche populations is unparalleled in the market research world. The company offers a variety of market research tools integrated into its in-home usage testing (IHUT) platform, among these MaxDiff.

A simple yet powerful tool in your market research toolbox

MaxDiff is an excellent way to see how your customers value certain features relative to one another, and it can complement a variety of other research methods like sensory testing for a holistic approach. It’s best when used as part of a broader research strategy rather than a standalone method.

With the Highlight platform, survey modules are built by product research experts–that means every Highlight user can be a product research pro. Easy audience building, live data collection, and instant statistical analysis gives product development teams the insights they need for product launch (or relaunch) success.