Conjoint analysis is a form of statistical analysis that firms use in market research to understand how customers value different components or features of their products or services.
In other words it's a research technique to measure what consumers value most about your products and services. For example, a TV manufacturer would want to know if customers value picture or sound quality more, or if they value low price more than picture quality. Conjoint analysis helps put a value on each feature, allowing you to tailor your products and services to what most consumers are seeking.
In this case, we will simulate data obtained from a crisp retail vendor that has asked some of its customers to rank its products according to their level of preference.
In this case, its can be see that there are factors that are positively related, such as the weight of 100 grams, and the option for the product to be fat-free, while there are other factors that are negatively related, such as the weight of 400 grams, followed by that the product is not fat-free.
We have been able uncover consumer preferences that otherwise would be difficult to determine. We used OLS to estimate the performance of different combinations of features and try to isolate the impact of each of the possible configurations in the overall perception of the client to uncover where the consumer perceives the value.
This has allowed us to create an overview of the factors that drive users to buy a product, and even be able to predict how a new combination of features will perform by using ML algorithms.