Optimizing marketing spend and maximizing profit through data insights
Data Cleaning: We removed missing and unreliable values, grouping channel observations by date to ensure data integrity.
Initial Regression Results: The estimated non-linear model shows how ad spend affects GP2, revealing Q4 overrepresentation due to holiday shopping.
Excluding Q3 & Q4: The model adjusted to show the expected relation between GP2 and cost, revealing under-investment in Q1 and Q2 marketing.
SEM COP Analysis: Weekly data reveals COP bandwidth for each quarter to maximize GP2, with different strategies for each season.
Conclusions: We recommend increasing marketing spend in Q1-Q3 to reach optimal GP2, while also adjusting for specific channels like SEM.