Correspondence analysis

Condenses multi-variable categorical relationships onto a two-dimensional perceptual grid (CA).

Setup and display

• Plots Factor 1 (horizontal) against Factor 2 (vertical), aiming for at least 70% combined variance explained.
• Metrics: Percent Influence (% INF), Absolute Contribution (ABS), Relative Contribution (REL).
• Entity delineation: row variables and column variables are shown with distinct symbols and color markers, separating profiled entities (such as brands) from descriptive attributes (such as lifestyle indicators).
• Custom axis labels: rename Factor 1 and Factor 2 (for example "Value vs. Premium" and "Traditional vs. Modern") for presentation-ready maps.
• Annotation styling: Annotation Mode supports custom line and text annotation colors (shared with Explore - P-Maps).

Use cases and benefits

• Segmentation and targeting: identify audience segments and their preferences, for example from media consumption habits, to fine-tune targeting.
• Media channel optimization: see how media platforms relate to consumer responses or brand metrics to allocate budget.
• Competitor analysis: like Cluster Analysis, cluster brands and products into competitive sets; brands sharing a factor compete more with each other.
• Identifying psychographic drivers: run before a Cluster Analysis to find which psychographic variables have the strongest associations with audience segments and products.

Interpreting the map

• Factors: typically Factor 1 on the X axis (left to right) and Factor 2 on the Y axis (top to bottom).
• Proximity: items close together are very similar (for example brands or age groups).
• Relationships between items: to relate active rows and columns (for example a brand and a demographic), draw lines to the origin (zero). Obtuse angles indicate a negative correlation.
• Distance from center: the closer an item is to the center, the less the map says about it; the further from the middle, the more strongly it relates to something else on the map.
• Variance explained: a robust analysis reaches 70% or more across Factor 1 and Factor 2 combined.

Key metrics in the results table