Propensity models are a robust application of machine learning in marketing. These models use historical examples of customer behaviour to make predictions about future behaviour. The predictions generated by the propensity model are commonly used to know the likelihood of a customer purchasing a specific product or taking over a particular offer inside a given time-frame.
In essence, propensity models are examples of the machine learning technique often called classification. What makes propensity models unique is the issue statement they solve and the way the output must be crafted to be used in marketing.
The output of a propensity model is a probability rating describing the expected likelihood of the specified customer behaviour. This rating may be used to create customer segments or rank customers for increased personalisation and targeting of recent products or offers.
In this text, I’ll provide an end-to-end practical tutorial describing the right way to construct a propensity model ready to be used by a marketing team.
That is the primary in a series of hands-on Python tutorials I’ll be writing…
