Purpose

This analysis applies Shapley Value Attribution to quantify how marketing channels and campaigns contribute to revenue generation across 2,205 customers. The objective is to identify which channel mix and campaign combinations drive the highest-value customer segments, enabling data-driven budget allocation decisions.

Key Findings

• Store Channel Dominance: Generates $443,213 (39% of total revenue) with highest revenue-per-purchase ($35), despite serving 2,195 customers
• Campaign 2 Attribution: Highest Shapley value (0.2) with 66.7% conversion rate, though smallest acceptance base (30 customers)
• Campaign 4 & 3 Negative Attribution: Shapley values of -0.38 and -0.29 respectively, indicating potential cannibalization or poor targeting
• Income Segment Correlation: High-income customers accept 0.74 campaigns on average vs. 0.08 for low-income, spending $1,150 vs. $61 average
• Overall Conversion: 15.1% baseline conversion rate across all channels; customers accept only 0.3 campaigns on average

Interpretation

The analysis reveals a clear performance hierarchy: Store and Web channels drive 66% of revenue, while Catalog and Deals contribute 34%. Campaign effectiveness varies dramatically—top performers

Purpose

This section documents the data preprocessing pipeline for the marketing attribution analysis, showing how raw customer and campaign data was prepared for Shapley Value attribution modeling. Perfect retention indicates no data loss during cleaning, which is critical for ensuring the attribution analysis covers the complete customer base and produces reliable revenue allocation insights across channels and campaigns.

Key Findings

• Retention Rate: 100% (2,205 rows preserved) - No observations were removed during data quality checks or mapping processes
• Rows Removed: 0 - The dataset required no filtering for missing values, duplicates, or invalid records
• Train/Test Split: N/A - Data was processed as a complete cohort rather than partitioned, appropriate for attribution analysis which requires full customer journey visibility

Interpretation

The perfect retention rate indicates exceptionally clean source data with no missing mapped values or quality issues requiring exclusion. This is favorable for attribution modeling, as the Shapley Value algorithm operates on the complete 2,205-customer population, ensuring that revenue attribution across all five campaigns and four channels reflects the true customer base without selection bias. The absence of train/test splits aligns with attribution analysis objectives, which prioritize comprehensive channel-mix evaluation over predictive generalization.

Context

While 100% retention suggests high data quality, this metric alone does not reveal potential issues in feature engineering, outlier handling, or data consistency across the

Purpose

This executive summary synthesizes the marketing attribution analysis across 2,205 customers to answer which channels and campaigns drive the most revenue. The analysis uses Shapley Value Attribution to fairly allocate credit across touchpoints, enabling data-driven budget allocation decisions.

Key Findings

• Campaign 2 Attribution: 20% of total attribution credit—highest performer despite smallest acceptance rate (1.4%), indicating exceptional conversion quality
• Store Channel Dominance: $443,213 revenue (39% of total) with highest customer penetration (2,195 users), establishing it as the primary revenue engine
• Campaign Efficiency Paradox: Campaigns 3 and 4 show negative Shapley values (-29%, -38%) despite similar acceptance rates, suggesting they cannibalize rather than drive incremental revenue
• Income Segment Concentration: High-income customers accept 0.74 campaigns on average versus 0.08 for low-income, generating 18.8x higher average spend ($1,150 vs. $61)
• Overall Conversion Rate: 15.1% baseline with significant variance by campaign (37.8%–66.7%), indicating substantial optimization potential

Interpretation

The analysis successfully identifies that revenue concentration follows a quality-over-quantity pattern. Campaign 2's outsized attribution despite low volume, combined with Store's

Purpose

This section applies Shapley value attribution to fairly distribute conversion credit across campaigns based on their actual contribution to customer decisions. Understanding which campaigns drive conversions—independent of scale—reveals which initiatives deserve budget prioritization to maximize revenue impact across the 2,205 customer base.

Key Findings

• Campaign 2 Shapley Value: 0.2 (highest) with 66.7% conversion rate—demonstrates strongest marginal contribution to conversions despite smallest acceptance volume (30 customers)
• Campaign 1 & 5 Performance: Shapley values of 0.17 and 0.15 respectively, with 55.6% and 56.5% conversion rates, showing consistent positive attribution
• Campaign 3 & 4 Negative Attribution: Shapley values of -0.29 and -0.38 indicate these campaigns detract from conversion likelihood, despite similar acceptance rates (7.4%)
• Low Campaign Penetration: Average of 0.3 campaigns accepted per customer reveals limited multi-campaign engagement across the base

Interpretation

Campaign 2's superior Shapley value reflects its disproportionate conversion efficiency—it converts 66.7% of exposed customers versus the 15.1% baseline. Conversely, Campaigns 3 and 4 show negative attribution, suggesting they may

Purpose

This section quantifies revenue contribution across four purchase channels to identify which distribution methods drive the most value. Understanding channel performance is critical for optimizing marketing spend allocation and identifying which channel mix produces the highest-value customers—a core objective of this attribution analysis.

Key Findings

• Store Channel: $443,213 revenue (39% of total) with 12,841 purchases—the dominant revenue driver despite serving 2,195 customers
• Web Channel: $307,262 revenue (27%) with 9,042 purchases and strong customer reach (2,163 users), demonstrating balanced volume and efficiency
• Catalog Channel: $262,003 revenue (23%) with the highest revenue-per-purchase ($45), indicating premium transaction value despite lowest customer base (1,634)
• Deals Channel: $124,265 revenue (10.9%) with lowest revenue-per-purchase ($24), suggesting lower-margin transactions despite broad customer participation (2,166 users)

Interpretation

Channel efficiency varies significantly: Store and Web generate consistent mid-range revenue per purchase ($34–$35), while Catalog commands premium pricing ($45/transaction) and Deals operates on volume with minimal margin. The Store channel's dominance reflects both high purchase frequency and broad customer adoption, making it the primary revenue engine. However, Catalog's efficiency suggests selective targeting

Purpose

This section reveals how campaign acceptance varies dramatically across customer income tiers, exposing a critical segmentation pattern in marketing effectiveness. Understanding these differences is essential for optimizing channel mix and campaign ROI, since high-value customers respond selectively while lower-income segments show minimal engagement across most campaigns.

Key Findings

• High-Income Acceptance Range: 14–27.6% across campaigns, with Campaign 5 achieving peak performance at 27.6%—nearly 14× higher than low-income segments
• Low-Income Acceptance: Consistently 0% across Campaigns 1, 2, and 5; only Campaign 3 and 4 show minimal engagement (0–0.7%)
• Segment Gradient: Acceptance increases monotonically from Low (0–0%) → Medium-Low (0–0.7%) → Medium-High (1.6–3.1%) → High (14–27.6%)
• Campaign Differentiation: Campaign 5 and Campaign 1 show strongest high-income performance; Campaigns 3–4 underperform across all segments

Interpretation

The data confirms the expected inverse relationship: high-income customers are selective responders but deliver disproportionate value when engaged, while low-income segments show near-zero campaign acceptance. This 27.6% peak

Purpose

This section identifies which product categories generate the most revenue and reveals customer spending patterns by category. Understanding category performance is essential for aligning marketing campaigns with high-value product segments and tailoring promotional content to drive revenue from premium versus volume-based categories.

Key Findings

• Wines Category: $675,093 total spend (59.4% of revenue) with $306.16 average spend per customer—a premium, high-value segment
• Meat Category: $364,513 total spend (32.1% of revenue) with $165.31 average spend—a mid-tier segment with solid volume
• Gold Category: $97,146 total spend (8.5% of revenue) with $44.06 average spend—a lower-priority segment with minimal customer value
• Spend Distribution: Wines dominates revenue despite representing only one of three categories, indicating highly concentrated value

Interpretation

Wines is the clear revenue driver, accounting for nearly 60% of total spend with the highest per-customer value. This premium segment should be the primary focus for high-income customer campaigns (particularly the High income segment, which averages $1,150 spend). Meat provides secondary revenue through moderate per-customer spending. Gold's minimal contribution suggests it either serves as an entry-level category or requires campaign repositioning to increase adoption among higher-income segments

Purpose

This section synthesizes attribution analysis findings to identify which marketing channels and campaigns drive the most revenue. It translates complex Shapley Value calculations into actionable insights about channel mix effectiveness and campaign performance, directly addressing the core business question: which marketing investments generate the highest-value customer outcomes.

Key Findings

• Campaign 2 Attribution: Shapley value of 0.2 (highest positive attribution) with 66.7% conversion rate, despite smallest acceptance volume (30 customers), indicating disproportionate revenue impact per engaged customer
• Store Channel Dominance: Generates $443,213 (39% of total revenue) across 12,841 purchases from 2,195 customers, establishing it as the primary revenue driver
• Negative Attribution Campaigns: Campaigns 3 and 4 show negative Shapley values (-0.29, -0.38), suggesting these reduce overall attribution value despite moderate conversion rates
• Income Segment Correlation: High-income customers accept 0.74 campaigns on average versus 0.08 for low-income, with 18.8x higher average spend ($1,150 vs. $61)

Interpretation

The analysis reveals that campaign effectiveness is not determined by volume alone—Campaign 2's small but highly-converting audience contributes more attributed value than larger campaigns. Store channel's domin

Purpose

This section segments customers by income level to reveal how campaign receptiveness, spending behavior, and engagement patterns vary across the customer base. Understanding these profiles is critical for identifying which segments drive revenue and which represent untapped or at-risk opportunities within the overall attribution analysis.

Key Findings

• High-Income Segment: 0.74 avg campaigns accepted with $1,150 avg spend—9x more engaged than Low-income customers, representing the most valuable and responsive audience
• Low-Income Segment: 0.08 avg campaigns accepted with only $61 avg spend—minimal engagement despite representing 25% of the customer base
• Recency Consistency: All segments show ~49-50 days average recency, indicating uniform purchase frequency across income levels despite vastly different acceptance rates
• Engagement Gradient: Clear linear progression from Low (0.08) → Medium-Low (0.14) → Medium-High (0.23) → High (0.74) campaigns accepted

Interpretation

Income is a strong predictor of campaign receptiveness and lifetime value. High-income customers accept campaigns at 9x the rate of low-income customers, directly explaining why certain campaigns (Campaign 2, 1, 5) achieve higher Shapley values—they likely resonate with affluent segments. The uniform recency across segments suggests purchase timing is consistent

Purpose

This section applies Shapley value attribution—a game-theory approach—to fairly distribute conversion credit across campaigns. Rather than assigning all credit to the "last click," Shapley values measure each campaign's average marginal contribution to conversions across all possible channel combinations. This directly addresses the core objective: identifying which campaigns and channels drive revenue and which customer segments respond best.

Key Findings

• Campaign 2 Shapley Value (0.2): Strongest positive contributor to conversions; delivers 20% of total attribution credit despite smallest acceptance rate (1.4%), indicating high-quality conversions
• Campaign 4 & 3 Negative Values (-0.38, -0.29): These campaigns suppress conversion rates when combined with others, suggesting potential audience overlap or message fatigue
• Positive vs. Negative Split: Three campaigns lift conversions (Campaigns 1, 2, 5); two actively detract, revealing fundamental differences in campaign effectiveness and synergy

Interpretation

Shapley values reveal that campaign quality varies dramatically—not by volume, but by conversion impact. Campaign 2's high Shapley value despite low acceptance suggests it targets high-intent segments or creates strong downstream effects. Conversely, negative values indicate campaigns that may cannibalize conversions or reach saturated audiences. These patterns align with the **income segment analysis