Surveys are invaluable for businesses. They provide feedback that customers and employees often feel uncomfortable providing in person. Surveys, forms of analysis, and other tools, if correctly used, allow businesses to be more aware and better off than their competition. MaxDiff analysis is one of many unique tools that businesses can harness in order to innovate and expand.
MaxDiff analysis is an investigative approach to measure survey respondents’ preferences for different attributes. MaxDiff survey questions list a set of at least two different attributes and ask respondents to choose among the attributes by ranking them as ‘most important’ and ‘least important’ or ‘best’ and ‘worst’.
To clarify, attributes are individual options or items that appear in a survey set. A set or data set is a group of attributes that researchers design in order to set attributes against one another. Sets require survey respondents to determine which attributes are most valuable to them.
MaxDiff surveys can be designed to include more than one set, each with different attributes, to provide broad research analysis. Greater amounts of data sets require respondents to compare more individual features against one another, providing data on a greater number of attributes.
MaxDiff analysis is a simpler version of conjoint analysis. The two different tools can be applied to similar situations, but MaxDiff is generally easier to use and more comprehensive for research analysis. With MaxDiff analysis, all attributes are measured on a common scale and can be directly compared. With conjoint analysis, respondents and survey researchers can only compare the utilities within each item.
MaxDiff survey questions have multiple applications including:
- Customer satisfaction
- Employee satisfaction
- Compliance
- Access Control
- Risk Management
- Asset Management
- Privacy
MaxDiff surveys can help businesses answer the following questions
- What is a customer focused on most when they make a purchase?
- How do different marketing tools and product claims resonate resonate with the target audience?
- How do consumers value different attributes when asked to choose from a list?
- What trade-offs would customers in order to be provided with other features, goods, or services?
- How do different brands compare to one another?
- What do employees value the most from their employer?
MaxDiff Analysis in Action
MaxDiff is beneficial in situations when businesses want to identify the preferences of consumers and employees. The survey questions are designed to provide several attributes and only two possible answers, requiring respondents to compare different attributes and choose what they value the most.
For example, a restaurant company might ask “what are the most and least important aspects of your restaurant experience?” The attributes listed can include food quality, polite service, affordable cost, and quick cooking time. Respondents can place ‘most important’ and ‘least important’ each on a single attribute.
If 8 out of 10 respondents place ‘most important’ on food quality and 6 place ‘least important’ on quick cooking time, the restaurant company can observe that they should focus more attention on cooking quality food than sacrificing food quality to make order turnaround as quick as possible. They will also be able to analyze the fact that polite service and affordable cost are important, but not the priority of the restaurant experience.
Survey designers can add more than one set of questions to require respondents to make more comparisons, either between unmentioned items or a mix of already-mentioned items from different data sets.
For example, the restaurant company above can add a new data set with food quality, polite service, restaurant decor, and restaurant location. If the restaurant location attribute receives the greatest amount of ‘most important’ answers, the restaurant company will understand that a restaurant’s location is of even greater importance than food quality.
MaxDiff Scoring
Survey researchers use the following formula to score and analyze the different attributes of a MaxDiff survey question: the number of times an attribute was selected as ‘best’ or ‘most important’ subtracted by the number of times an attribute was selected as ‘worst’ or ‘least important’, divided by the number of times the attribute appeared.
For example, if 12 survey respondents choose food quality as the ‘most important’ attribute between the two data sets in which the attribute appears and ‘least important’ 4 times, divided by the 20 times the attribute appears, the MaxDiff score would be 0.40. If only 2 survey respondents choose quick cooking time as the ‘most important’ attribute only twice in the one data set in which the attribute appeared, and ‘least important’ 6 times, the MaxDiff score would be -0.40.
A higher score indicates that the attribute is appealing to respondents. A positive score indicates that the attribute was selected as ‘best’ more often than ‘worst’. A negative score indicates that the opposite is true. If an attribute score is 0, it was either chosen as ‘best’ and ‘worst’ an equal amount, or the attribute was never chosen.
Why MaxDiff?
Using a sufficient sample size, usually 150 at the bare minimum, businesses and organizations can harness MaxDiff analysis as a tool to improve their goods, services, and other integral aspects of their business. While other survey tools require respondents to prioritize and choose attributes of the highest value, MaxDiff is perhaps the most widely respected and least flawed of these tools.
Other rating questions are susceptible to scale meaning bias, scale use bias, and lack discrimination. Surveys that rank attributes as ‘important’, ‘very important’, and ‘most important’ are less accurate because different respondents have different measures of what makes something ‘very important’ or ‘important’. With MaxDiff survey questions, respondents only have to choose ‘most important’ and ‘least important’, providing no possibility for scale use bias. MaxDiff questions also show greater discrimination among different attributes and between responses than rating scale questions.
MaxDiff surveys are also easier to analyze and understand than most other rating survey types. Respondents are rarely confused with the format of MaxDiff questions, which positively affects survey accuracy. Furthermore, the formula for analyzing MaxDiff scores is relatively simple. MaxDiff analysis is a valuable tool for businesses. It is only by understanding consumers, employees, and the market that businesses can improve and rank above their competition.
