From the course: Marketing Attribution and Mix Modeling
Unlock the full course today
Join today to access over 23,200 courses taught by industry experts.
Feature transformation with diminishing returns and adstocks
From the course: Marketing Attribution and Mix Modeling
Feature transformation with diminishing returns and adstocks
- [Instructor] Not all variables have a linear relationship with sales. For example, brand marketing tends to have a lagged impact long past when the ad ran. And most advertising gets less efficient at higher spend levels. Let's work through how to transform the nonlinear variables to use in marketing mix models. This marketing mix model has already built using the LINEST function here. It's for an ice cream store and we want to know whether there is a lagged effect and also a diminishing returns effect. So, we're going to work through how to do this and build some intuition for what that means. Okay. So we're going to walk through adstocks first. So to add adstocks, I think it's quite helpful just to make an adstock cell up here and then zero is the default adstock. Then in order to make adstocks work, we need to multiply the data by this adstock, right? So in the first cell, we can see we're referencing data C2, and…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
(Locked)
Before and after an event: Trend analysis4m 40s
-
(Locked)
Linear regression with a single variable6m 4s
-
(Locked)
Variables with positive and negative correlations5m 45s
-
(Locked)
Multivariable regression: Building your marketing mix model11m 30s
-
(Locked)
Feature transformation with diminishing returns and adstocks10m 27s
-
(Locked)
Statistical tests to validate your model's accuracy12m 47s
-
(Locked)
Forecasting future scenarios for planning11m 25s
-
(Locked)
-
-