How we applied Structural Equation Modeling to a beverage brand in North America

One challenge posed by the high number of variables tracked in the non-alcoholic beverage industry is the ability to prioritize which variable types (e.g., imageries vs occasions, vs motivations) and which variables are most likely to influence consumer behavior. This can at times result in too much focus on a few “silver bullets” or analysis paralysis

What is the actual path to teen recruitment for a brand?

Which variables drive adoption?

Which drivers come first and how are they connected?

• Use factor analysis to identify groups of variables that fit together
• Apply structural equation modeling to test alternative causal models of how various drivers (e.g., imageries, reasons for drinking, etc.) and other information tracked (e.g., occasions, packs, etc.) drive recruitment
• Collaborative optimization & refinement; hypothesis based testing to maximize relevance & actionability

• Brand Affinity and Favorite Brand have strong direct relationships with incidence
• Fun & enjoyment perceptions have an indirect relationship with incidence, but are the most impactful overall
• Brand imageries are stronger than category imageries except when it comes to health
• Client kicked off more qualitative research to extract deeper consumer insights on some of the imageries that had been identified as main drivers. The new insights led to a shift away from a Taste-only focus, and to a new advertising campaign