“Our Consumer Decision Tree (CDT) research is helping Kellogg’s understand what influences consumer choice as well as giving us guidance about major brands and segments and where they fit within the category structure.”, says Tiernan Summins, Sr. Manager, Market Insights, “But we’re not relying on your mother’s CDT; we’ve put a new twist on this old idea.”

Historically, CPG companies have either used scanner data (market basket analysis) or qualitative research to create CDTs.  “Each method had its strengths and weaknesses but ultimately we felt like there must be a better way,” explained Summins. 

Old Way

Using scanner data allows for a large sample of respondents, taking into account many unique variations in shopping habits and personal tendencies, but is overly reliant on the objective characteristics of products as driving purchase decisions and a failure to understand the reasons underlying the decisions consumers make about what to purchase.
The qualitative approach provides a richer understanding of the shoppers’ behavior and provides hypotheses about the impact of shelf arrangement, sub-category relationships and other key retail decisions, but this approach cannot be generalized to the larger population and is highly dependent on the quality of the interviewers. 
Both methods are highly influenced by the structure placed on the category; the results can seem to be a self-fulfilling prophecy based on the inputs to the research process.

New Way

Decision Insight’s process of starting with shopping breaks the mold of self-fulfilling results and brings truly innovative thinking.   The hybrid quant/qual solution measures actual consumer behavior (through simulated shopping) followed by qualitative follow-up that expresses how shoppers make decisions and identifies the criteria they use to evaluate the products.  Ultimately, we learn what drives consumer choice and how consumers understand the interplay among the products in the category.  In addition, we can pinpoint how retail channel influences the decision process by creating unique trees for each channel.