The team should pay attention to imbalanced classes, using metrics like precision and recall rather than just accuracy. They should also look at the ROC curve rather than just the single threshold. Sokolova (2009) provides guidance on selecting appropriate metrics based on factors like class balance and error costs. The key steps the team should follow are: 1) Split the available data into training and holdout test sets. 2) Train the model on the training data. 3) Make predictions on the holdout set using the trained model. 4) Calculate evaluation metrics like accuracy, precision, recall, F1 score, and ROC curve on the test set predictions. 5) Tune the model and thresholds to optimize performance. 6) Retrain on all data once the model is finalized