How to build an acquisition models?

Question

I have historical transaction data for a B2B business. I want to prepare acquisition models where the model will predict the potential buyers who haven't purchased from the business yet. In addition to the transaction data, I have customer data where I have the details about the different businesses in the region (Their Sales, Profitability, Size, Location, and Industry) from an external source.

How do I translate this business problem into a data science problem? What should be my approach in building these models? Are there any resources available for similar problems? I really appreciate any help you can provide.

Answer 1

You have a binary classification problem. In other words you want a model to decide if a user will be customer or no. One way is to create two sets of rows. One with users that became customers -1 and a second with users that didn't - 0. This column with values 0 or 1 is the Y, aka truth label. You need historical data with the features you think play some role ie how many clicks a user did in a website or his location/age. These features are what an algorithm will use to adjust its parameters - what we call training. Once you train a model you need to see how good it is. A simple but effective way for such problem is to use write python program that uses the xgboost or scikit-learn libraries. There are thousands of tutorials with for similar methods. I hope this gave you some context to search, good luck.