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

For questions about this case study contact Nick Lucas.

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, occasions, or motivations) and which specific 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 brand adoption? Which drivers come first and how are they connected?

• Use statistical 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.
• Collaborate with client on optimization and refinement including hypothesis-based testing to maximize relevance and actionability for marketing campaigns.

• Brand Affinity and “Favorite Brand” both have strong direct relationships with incidence.
• Fun and 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.
• The 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.