One of the problems in asking a person directly what features or attributes of a product contributed to the purchase is that their responses, though truthful, may not accurately reflect the contribution a feature makes to the purchase process. A decision is usually a trade-off between multiple features, which can be hard to capture in direct questioning of the appeal of individual elements.
In addition, sometimes respondents give socially acceptable answers and display other biases which may lead to a difference between claimed behaviour and actual behaviour.
Through Choice Modelling an offer is broken down into its individual “attributes” (i.e. price, options, features) and different combinations (packages) are generated. Through statistical estimation, we are able to identify the contribution of each of the individual attributes to consumer choice. It is well suited for estimating market demand, price strategies, predict willingness to pay and estimate market share.
Choice modelling outputs can also be used to generate market share scenarios by adjusting product features and attributes. It also allows the researcher to look at the interactions between any set of attributes, such as price and brand.