Classification model lifecycle and click-through prediction

Walk through every step of building a classification model and the techniques at each step: framing and labels, data collection and cleaning, features, model choice, training, evaluation, deployment and monitoring. Then discuss predicting click-through rate with a binary model, and how you would find out why its performance dropped in production.