Discrete Choice Versus Decision Tree
Determining the configuration of a new product or service is a major responsibility of market researchers with enormous potential effects on return-on-investment (ROI). Given the importance of these decisions, it is not surprising that a number forms a utility score.
- Conjoint Analysis - These models let consumers evaluate a series of real or hypothetical products or services that are defined by the attributes they feature. The responses of research participants are used to identify a relative value of each feature.
- Adapted Conjoint - This model of conjoint analysis facilitates the analysis of a very large number of product or service attributes, or levels of attributes.
- Decision Tree - These models are used in market research to represent the process of decision making, which can include outcomes that were a result of chance, resource availability, or utility.
Bridge the Gap Between Insight and Optimization: Purchase Decision Hierarchy
The research that proceeds a product launch must dovetail with many levels of information. Consideration of ways to optimize a service or product line will tend to dominate the earliest phases of the run-up to a product launch, but investigating the decision processes that consumers put into play at the point-of-purchase can help to shape those early considerations. A hierarchy of sorts engages consumers in their purchase decisions. This hierarchy comes most easily into focus when a variety of sources of data and information are utilized, including - most importantly - marketing research and sales data.
While sales data can be helpful with regard to insights into weakening performance or declining market share, it doesn't have much predictive capacity. More intimate customer knowledge can provide insights into what is likely to happen to market share when a product is temporarily out-of-stock or removed from a product line. Market research can provide these types of insights, as well as an understanding of new product preference share or about switching behavior from existing products to a newly launched product.
Product or service optimization can be a costly endeavor and is invariably a high-risk option that demands the highest levels of precision and the capacity for broad and deep scenario simulation. Both discrete choice analysis (DCA) or choice-based conjoint (CBC) processes can meet these market research demands.
Decision Trees: a Budget-Conscious Option
Decision tree models can be used to develop a deeper understanding of consumers' hierarchical purchase behavior. Learning what product or service attributes trump one another and how, for instance, these dynamics relate to the shelf organization in bricks and mortar environments, puts a fine point on consumer insight. Decision tree models can be manipulated to focus on either brand perspectives or product perspectives. Decision tree models often capitalize on a visual representation of the products being considered in order to facilitate the research process.
The construction of a decision tree is central to its ability to elicit and capture hierarchical responses from consumers within the context of an intuitive survey experience. Because of the pivotal nature of decision tree market research to important marketing direction-setting, decision tree methods must have structural integrity and confidently reduce respondent burden. Going the extra mile in the design of decision tree market research will help to avoid the pitfalls that surveys research can encounter.
The effect of speedster respondents on the final surveys research outcomes can have a substantively negative impact on associated business decisions. It is important to have a data quality cleaning process that identifies speedster respondents and removes their data from the dataset. For these reasons, market researchers may employ a verification process that is built into the survey's research or entails a follow-up opportunity with each respondent. Those survey responses can be reviewed and adjusted as needed.