Marketing Campaign A/B Testing Analysis
Analyzed marketing campaign data to determine the most effective promotion strategy across market sizes and store age groups using statistical testing and data visualization.
Overview
This project involved an analysis of a fast food marketing A/B testing dataset. The goal was to evaluate the effectiveness of three promotional strategies across different market sizes (Small, Medium, Large) and varying store ages (1–28 years). The analysis was conducted using Python, supported by Pandas for data wrangling, Seaborn and Matplotlib for visualization, and Scipy/Statsmodels for statistical validation.
Key objectives included identifying the best-performing promotion, understanding performance consistency over time, and evaluating promotion effectiveness across store demographics.
Challenges
- Adjusting comparisons for varied market sizes and store age ranges
- Statistically validating differences between promotions to support business decisions
- Tracking temporal performance across a four-week campaign period
Solution
To address these challenges, I employed a structured analysis pipeline:
- Segmentation: Grouped data by market size and store age bins for fair comparative analysis
- Statistical Testing: Conducted ANOVA to detect overall differences and used Tukey HSD for detailed pairwise comparisons
- Trend Analysis: Visualized weekly sales trends to assess performance consistency
Results
The analysis revealed clear insights:
- Promotion 1: Outperformed others with the highest average sales (58.1K), consistent performance across all segments, and strong results in 11–15-year-old stores
- Promotion 2: Consistently underperformed and is recommended for discontinuation
- Promotion 3: Delivered moderate performance with a slight downward trend