Analysis of 2216 customers identified 4 distinct behavioral personas. The 'Premium' segment is the highest-value group, averaging $1310 in total spend — 783% more than the lowest-value segment. K-means clustering explains 57.5% of total customer variance, confirming that these personas are statistically well-separated and actionable for targeting.

Each row represents one customer persona with its average income, recency, and category spend. The 'Premium' segment has the highest average income ($74851) and wine spend ($537 over two years), making it the clearest premium target. Recency scores show how recently each segment purchased — lower values indicate more active customers.

The 'Premium' segment outspends all others, averaging $1310 across all product categories over two years. The gap between the top and bottom segments is $1162 — a 783% premium that justifies differentiated acquisition budgets per persona.

The box plot reveals how cleanly household income separates customer segments. Across 2216 customers, the median income is $51382 but varies substantially by cluster. High overlap between cluster income ranges indicates that spending behavior — not income alone — drives cluster membership, underscoring the value of the multi-feature approach.

The heatmap shows Pearson correlations between all spend categories. The strongest cross-category link is between Mnt Fish Products and Mnt Sweet Products (r = 0.58), suggesting these products are frequently bought by the same customers and form a natural bundle. Highly correlated pairs identify cross-sell opportunities where promoting one category to buyers of the other is likely to find a receptive audience.

Channel preferences differ markedly across segments, informing where to place promotional spend. The 'Tier 1' segment leads in in-store purchases, while other segments may prefer web or catalog channels. Matching campaign delivery channel to segment preference improves conversion rates and reduces wasted impressions.

Each point is one customer; color indicates cluster membership. A clear diagonal separation between premium and budget clusters confirms that income and wine spend together are strong discriminators. Customers clustered in the upper-right quadrant (high income, high wine spend) represent the highest-lifetime-value cohort and the most viable target for premium campaigns.

Campaign acceptance rates show which segments are most responsive to promotional offers. The 'Premium' segment accepts campaigns at a 45% rate — 28.6 percentage points above the least responsive segment. Concentrating campaign budget on high-acceptance segments delivers better ROI than blanket campaigns across the entire customer base.