Start from editor screen/dataset details screen and click 'Select use case' button.
Select 'Customer lifetime value' use case.
For transaction amount, choose the column that indicates how much your user spends on a single transaction. Please note: when you click on the input, it automatically filters out the columns that are not a number column type.
For transaction date, chose the column that has the date of transaction. When you click on the input, only column with datetime type will appear.
Performance window is how far out you want the model to predict. If you are unsure of which performance window to use, then you should choose the recommended window that Mage automatically calculates based on the transaction date selected.
If you don't choose the recommended window, you should keep the window under a year because the farther out the model predicts, the less accurate the model predictions will be.
ID column is the column that contains the unique identifier for the user/customer you are trying to predict LTV for.
Under advanced options (optional)
Observation window: For best model performance results, we recommend using at least 3 months. The more data you can provide, the better your model may perform.
Please enter your model name and click on 'Build and train model.'
This will take some time but you will receive an email when the model is done training.
