Model Builder
Find the drivers of purchase intent
Build a logistic regression of purchase intent for the test product. Pick predictor variables — demographics, beer-drinking habits, brand awareness, package reactions — and see which ones move the headline.
The outcome is a top-2-box recode of Q21: 1 if a respondent rated purchase intent Extremely or Very likely; 0 otherwise. Baseline rate is 54.7% in the unmodified sample.
How to Use This Tool
The published model is loaded — six package-reaction questions are pre-selected. Click Run Analysis at the bottom to fit it, or add and remove variables to build your own version.
How to Use This Tool
The published model is loaded — six package-reaction questions are pre-selected. Click Run Analysis at the bottom to fit it, or add and remove variables to build your own version.
1. Outcome
The outcome is fixed for this case study: Purchase Intent (Top-2-Box on Q21). 1 = Extremely or Very likely; 0 = Somewhat / Slightly / Not at all likely.
2. Pick predictors
Six package-reaction questions are pre-selected — the published model. Use the search box and category buttons (Package Reaction, Brand Knowledge, Beer Preferences, Behavior, Demographics) to find others to add or to swap in.
3. Fit and read
Click Run Analysis to fit the model with whatever variables you've checked. Or use Auto-Build Standard to let the algorithm search all variables automatically, or Auto-Build Actionable Predictors to weight selection toward levers you can move through package design or marketing over fixed context like demographics.