Conjoint Analysis is a type of analysis based on a Conjoint module in a quantitative survey. The Conjoint module asks respondents assess their interest in a wide range of variations on a particular product or service offering. The analysis of this data allows researchers to identify which product or service offering will have the widest share of preference among consumers. This often includes data on how the ideal product should be priced. However, it is important to note that this ideal product/price will not necessarily result in the highest profit or maximize revenue. The main goal of Conjoint Analysis is to identify which product or service offering will have the widest appeal, not the one that will result in the largest amount of revenue.
There are 3 main types of Conjoint Analysis:
A company that specializes in pre-packaged health food is developing a new product in its line of snack bars, and they want to ensure its widespread appeal among consumers. First, they consider all the types of attributes that will be important factors when consumers consider purchasing the bar. This is their final attribute list:
Each of these attributes have different levels. The research team won’t be able to test all possible levels of the attributes, but they’re able to narrow down to these attribute level lists:
Now that they have their attributes and levels, they’re ready to design the Conjoint module. They decide to use the most popular kind of conjoint exercise, a Choice-Based Conjoint. To move forward with their research project, they buy a software subscription that will both design the module and analyze the output for them. First, the software uses an algorithm to create the product combinations that will be shown to respondents. It is too exhausting for respondents to analyze every possible combination of attribute levels, so the software creates a design that maximizes the possible combinations while minimizing the number of products respondents must assess.
Then, the team programs the survey to show the product combinations provided by the software. They include a series of screening questions to ensure that they reach the correct audience, and then add some demographic and attitudinal questions so they can do more in-depth analysis of their results.
Once they finish fielding the survey and they have their final dataset, they feed the conjoint data into their conjoint analysis software. The software contains an algorithm that looks at the data for the feature combinations that were shown, and uses that data to fill in the gaps for every possible feature combination that was not shown. Thus, it creates model for preference for all possible combinations of attribute levels. Using this model, a market simulator (either in Excel or Powerpoint) can be generated to show how share of preference changes when attribute levels are modified.
After this market simulator has been created, the team discovers that the bar with the widest share of preference is:
However, the team knows that they can’t make a high-protein peanut butter bar when the price and calorie count are that low. The profit margin for this product would be too low, so they must choose a different product. They try different attribute level combinations and discover that bars that are high-protein and lower-calorie are more popular, so they use the simulator to see what the next best choice would be.
Finally, they decide upon the product below, and enjoy its widespread popularity when it is launched nationally.