Advertising Performance Based on Personalization Breadth and Depth

The dataset comprises one year of advertising performance data from Meta Ads, linked with Google Analytics 4 tracking for an e-commerce retailer. Observations are aggregated at the campaign–week level, covering 24 distinct campaigns across multiple funnel stages (TOFU, MOFU, BOFU). Each record contains structural parameters (funnel stage, audience type, retargeting logic, creative personalization depth), behavioral metrics (impressions, click-through rate, conversions, revenue), and cost indicators (advertising spend, ROAS). Audience segmentation includes custom, lookalike, and broad targeting groups, with creative personalization classified as low-complexity or dynamic. Data preprocessing involved removal of incomplete records, normalization of numerical variables, and encoding of categorical features for modeling. The final dataset supports regression modeling, feature importance analysis, and unsupervised clustering to explore the relationship between personalization width/depth and advertising efficiency.