Have 1 year sales funnel data and want to predict what is the likelihood of winning a open deal this quarter (in next 90 days) or moving out to future quarters. This is a big problem in sales organizations where quarterly funnel predictions are based on a sales reps given date.
1: Use Random Forest/GBM to develop a classification model to predict Pr(win) a deal basis past historical data.
2: Predict if it will close in next 90 days
What kind of model shall I fit to predict on a time period. Also, found some articles which suggest running a Poisson regression or Poisson duration model. However, unable to understand how shall I create the training dataset to predict the time period. Please help me here.