Indeed Engineering is a brand that focuses on providing innovative solutions to enhance the user experience and improve business metrics. With a team dedicated to optimizing the Indeed Interview platform, Indeed Engineering aims to increase the number of interviews happening on the platform. To achieve this goal, Indeed Engineering utilizes SHAP plots, a statistical model, to identify impactful A/B test ideas that can significantly boost product growth.

By analyzing employer behavior on the employer dashboard and predicting interview invitations, Indeed Engineering creates a dataset that represents the number of times an employer clicks on specific UI elements. Using a tree-based model called CatBoost, Indeed Engineering trains the model to make predictions based on the dataset. The powerful combination of SHAP plot and the tree-based model provides both performance and interpretability.

By interpreting the SHAP plot, Indeed Engineering can identify positive and negative predictors that influence the targeted action of clicking the Set up interview button. With the insights gained from the SHAP plot, Indeed Engineering generates A/B test ideas to optimize the user interface. Positive predictors are made more prominent to boost target usage, while negative predictors are strategically diverted to increase target usage.

These test ideas have led to a significant increase in key business metrics, with a 5-10% improvement observed. Indeed Engineering's methodology not only provides valuable insights into user behavior but also enhances the overall user experience by focusing on features that have the most impact. By leveraging this methodology, businesses can make data-driven decisions, optimize their UI, and achieve better outcomes