Tuning a model is a core element in data science and requires experience. An integral part of the tuning process is the feature selection. We would like to suggest a simple process – calculate feature contribution data (or explainability) for different experiments, and then analyze the data using analytic tools. You learn about your features and choose the right ones for you. That will lead to better results, and easier and faster to achieve them. We will explain the process and show examples.