By: Bart Baesens, Seppe vanden Broucke
This QA first appeared in Data Science Briefings, the DataMiningApps newsletter. Also want to submit your question? Just Tweet us @DataMiningApps. Want to remain anonymous? Then send us a direct message and we’ll keep all your details private. Subscribe now for free if you want to be the first to receive our articles and stay up to data on data science news, or follow us @DataMiningApps.
You asked: Can you give a practical example of how a profit-driven perspective is different from a statistical perspective on analytics?
Our answer:
Let’s take the example of customer churn prediction. Many organizations nowadays develop and operate customer churn prediction models, allowing them to select and target the customers which are most likely to churn in a retention campaign. Adopting a profit-driven perspective in developing, evaluating and operating a customer churn prediction model means that you account for the value of customers. The higher the value of a customer, the more important it is to accurately estimate the risk of churn. Profit-driven predictive analytics allow you to do so. For evaluating the performance of a customer churn prediction model, we argument it is the profit that should be assessed which the model potentially generates when using it for selecting customers to include in a retention campaign. That is what profit-driven evaluation measures allow to do. Uplift modeling, finally, allows to further refine the analytical model setup by estimating the net effect of targeting a customer in a retention campaign, in terms of the net decrease in probability to churn. This allows to select so-called persuadables, meaning customers which can be persuaded by the campaign, rather than lost causes, which are customers who already made up their mind and will churn even when offered an incentive to remain loyal. As such, the return of marketing efforts can be optimized and significantly increased, as several case studies show.
