Predictive Analytics in Commerce

Chapters 6 › Unit 6: More Personal to the Customer View instructions Hide instructions

More Personal to the Customer

Step 1. Build the best possible churn model

It is time to create your own churn model. Start by creating a random forest on a subset of the data, and then (just like Nienke recommends) build a random forest on the whole dataset (be aware that this might be more time-consuming).

Do you get similar results when you use your whole dataset? And do the results seem logical to you?

Step 2. Elevator pitch

Set up a short presentation (an elevator pitch) to convince your boss to set up a retention campaign based on the churn model you just built. Mention the work you have done, what problem you are trying to solve and how you think you can solve it.

Need a little guidance on how to setup a good elevator pitch? Watch this movie and you are ready to go:

Churn Model


I ran a churn model on the full dataset which indeed was more time-consuming, but at the same time allowed for some heavy coffee-drinking.

Running the model on the full dataset, I identified that the model was stabilizing at around 350 trees while the lowest out-of-bag error was 8 mtry. This model had an r-squared of 58,95% and an AUC of 92,1%.

I however wanted to build a more simple but still accurate model, why I choose the best performing variables across the relative influence and node purity. This resulted in a model of only 8 variables, 350 trees and 8 mtry. This model had an r-squared of 56,81% and an AUC of 91,2%.

Overall the two models didn't seem to differ too much in their performance and investigating the overall accuracy, sensitivity, specificity and precision only supported that where I didn't find too signficiant deviations when comparing the models.

I have presented the elevator pitch as the first slide in my presentation and then the results from my simple model in the remaining slides.

enter image description here

Your Comment

Please login to leave a comment.