Introduction
Measuring marketing performance
()
1. Multi-touch Attribution Models
Last-click attribution: The default model
()
Time decay and conversion lags
()
Linear attribution: Treating all touches equally
()
First-click models: From awareness to acquisition
()
Position-based models and assigning credit
()
Data-driven attribution and machine learning
()
Click windows and view-through conversions
()
2. Marketing Mix Modeling
Before and after an event: Trend analysis
()
Linear regression with a single variable
()
Variables with positive and negative correlations
()
Multivariable regression: Building your marketing mix model
()
Feature transformation with diminishing returns and adstocks
()
Statistical tests to validate your model's accuracy
()
Forecasting future scenarios for planning
()
3. Incrementality and A/B Testing
A/B testing for statistical significance
()
Bandit testing: Optimizing for results over accuracy
()
Geo and lift testing to prove incrementality
()
"How did you hear about us?": Surveys and panel studies
()
Working with multiple attribution methods
()
Conclusion
Continuing to improve your model accuracy
()
Ex_Files_Marketing_Attribution_Mix_Modeling.zip
(225 KB)