40% improved accuracy in forecasting for package service provider

Customer satisfaction is the key to success for businesses. Our client, a leading package service provider with 600+ operating facilities, was undergoing the challenge of predicting the volume of ground packages by category. Our client needed a strategic partner that could create a reliable and efficient process for developing, deploying, and managing machine learning (ML) models and resolve the lack of collaboration among teams. Ascendion stepped up to help improve delivery excellence, customer experience, and driver safety, by optimizing the logistics for package deliveries.

Revolutionizing deliveries with real-time route genius

Ascendion’s experience in deep data engineering expertise (DataOps and MLOps) helped us leverage modern data strategies and drive technology expertise. We used Ascendion AVA ML Optimization studio and implemented Databricks MLOps with a MLFlow solution. Our solution included using forecast models such as ARIMA, Prophet, exponential smoothing, and XGBoost within a unified framework.

Ascendion deployed MLFlow’s tracking capabilities to enable easy comparison and evaluation of the forecast models. With monitoring and logging mechanisms implemented, we tracked the performance of the deployed models.

Our solution provided automation and continuous integration to refine and enhance the models (based on user feedback and factors such as weather and traffic), optimize resource allocation and inventory, identify bottlenecks, and reduce delivery delays, resulting in improved customer satisfaction.

The results:

• Reduced deployment time by 65%
• Increased volume forecast accuracy by 40%
• 80% insights gain for better resource allocation and route planning
• 40% improvement in delivery times
• Real time response for proactive decision making