Predictive models are based on the premise that if X happens, Y will happen. In other words, based on what happened, predictive models try to predict what will happen.
Why did it happen? What could happen?
Let’s review the case of Target. Andrew Pole is a Data Scientist for the chain who, by analyzing the purchasing patterns of 25 products, was able to create a predictive model. In this way he was able to determine, with an accuracy of more than 80%, if the consumers were pregnant and what month they were going to give birth.
The goal –
To identify pregnant consumers and their shopping habits to understand what products Target should offer them.
The gathered data –
Target membership number and purchase history. With them, the data model was programmed to isolate female consumers (or associated couples) who began to buy products such as calcium, zinc, odourless lotions, disinfectants or much larger bags.
The moment in which they began to buy each of these products also coincided with the months of pregnancy, so the final data model assigned a “natal percentage” to each user, with 100% being the estimated time of delivery.
What do you think about this? Did you know this case?
You may also be interested in:
If you want to contact us directly, send us a message.