Penalty Analysis quantifies the impact of product attributes that are "Not Just About Right" (JAR) on Overall Liking.
The Logic
For each attribute (e.g., Sweetness), we group respondents into three buckets based on their JAR rating:
- Too Low (e.g., 1-2 on 5-pt scale)
- Just About Right (JAR) (e.g., 3 on 5-pt scale)
- Too High (e.g., 4-5 on 5-pt scale)
Calculations
1. Mean Drop
How much lower is the Overall Liking for the "Too Low" group compared to the "JAR" group?Mean Drop = Mean(JAR) - Mean(Too Low)
2. Weighted Penalty
A large drop matters less if only 1% of people feel that way. The Weighted Penalty combines the severity of the drop with the percentage of people affected.Weighted Penalty = Mean Drop × % of Respondents
Interpretation
We plot these on a chart:
- X-Axis: % of Respondents (How many people complain?)
- Y-Axis: Mean Drop (How much do they hate it?)
Attributes in the top-right corner (High %, High Drop) are your critical issues to fix.