MaxDiff is a statistical analysis widely used in market research to better understand what relative importance customers or consumers of a brand give to different attributes in a list. For example, what is more important to them when buying a mobile device: is screen size or storage capacity more important, camera quality or price? If we ask consumers, they will decide that everything is important and that they can’t do without anything, of course. The MaxDiff analysis forces them to choose between the different features the device may have and thus identify the combination they like the most.
Today we tell you how to apply the MaxDiff method step by step.
MaxDiff (abbreviation of Maximum Difference Scaling) is a statistical analysis that is frequently used to understand the importance that consumers give to each of the attributes that a brand or product may have. Although mostly used to prioritise product features, MaxDiff analysis also allows you to assess the relative importance of different marketing messages, or any list of items where it is necessary to understand which are the most and least important to the target audience.
In a MaxDiff analysis, respondents are presented with groups of items and asked to select the one they consider most important or attractive (maximum) and the least important or attractive (minimum). This is repeated several times with different combinations of items until each item has been indirectly compared with several other items in different combinations.
In the mobile phone example, the MaxDiff would be used by asking each participant to select, from a limited set of features (e.g. four), which they consider most important and which they consider least important. Through successive questions, they would be presented with new lists of four features to force their choice. The accumulation of partial choices allows the construction of a preference scale for each attribute.
MaxDiff analysis has some similarities with conjoint analysis, but they are not the same. Conjoint analysis is used to understand how different combinations of attributes affect the final choice. However, MaxDiff is used to measure the relative importance of a list of attributes or items, but not their combinations.
MaxDiff analysis is used when you want to ‘force’ the choice of different characteristics and the reason is that the quality of the information exceeds what is obtained with other types of questioning options:
In addition to these advantages, the MaxDiff questionnaire is easier for participants to answer, which may be important in reducing sample fatigue and improving questionnaire completion rates.
MaxDiff analysis can be applied whenever there is a need to understand the relative importance given by the target audience to different attributes. This need can arise in a variety of situations.
A MaxDiff questionnaire is organised in such a way that participants must choose, from a series of sets, which attribute they prefer and for which they have a lower preference. Let’s continue with the example of mobile phones. You could have a list of features like this:
Below, I show you how you could present MaxDiff’s questions in the quiz with some sample questions:
Question 1: Of the following attributes of a mobile phone, select the one you consider most important and the one you consider least important.
Question 2: Of the following attributes of a mobile phone, select the one you consider most important and the one you consider least important.
Question 3: Of the following attributes of a mobile phone, select the one you consider most important and the one you consider least important.
As you can see in the examples, in each question, the respondent sees a subset of features, in this case four. In a full questionnaire, the respondent would see multiple sets of four attributes at a time, covering all characteristics, and selecting the most important and least important in each set. This type of questionnaire is designed so that each characteristic is repeated in different sets and in different positions, allowing for a more robust analysis of preferences.
First, data collection is carried out through an online survey. Participants are asked to choose the attribute they consider most important and the least important from a subset of items, as we have shown in the example questionnaire. For each set presented, we can see directly which option was selected as the most important and which as the least important.
Statistical analysis is then carried out. You can do a simple analysis by counting the ‘most important’ and ‘least important’ selections for each attribute and calculating the net score by subtracting the number of times an attribute was selected as ‘least important’ from the total number of times it was selected as ‘most important’. For a more accurate analysis, you can employ advanced statistical models that allow for a more precise estimation of the relative importance of each attribute at the individual and group level. These analyses require the use of specialised software packages, or can be ordered from the We are testers support team.
Once the analysis is done, the preference scale is calculated. Each attribute is given a score and the results are converted into a common scale for easy interpretation. These scores are often normalised so that the average value is 0. Finally, MaxDiff scores are plotted in bar charts with the attributes ordered from highest to lowest preference. For further depth of analysis, you can do complementary analyses focusing on different customer profiles to understand if there is a difference in preferences between them.
The interpretation of the results allows you to determine the priority attributes to enhance and those of lesser importance that you can ‘sacrifice’ or use less intensively. In addition, this analysis allows us to see the distance of each attribute in relation to others. This helps to understand if any attribute stands out strongly or if several attributes have similar levels of importance.
The MaxDiff analysis is a very useful tool that you can use with the We are testers research platform. Our team of research experts will help you fine-tune the questionnaire to suit your needs and can perform the statistical analysis so you don’t have to worry about it. Contact us to find out all the details about how to apply MaxDiff analysis in your studies.
Update date 9 November, 2024
