SCENARIO Delivery Solution is a large logistics platform that provides last mile delivery services in the United States. They connect retailers with multiple delivery service providers, enabling them to choose the most cost-effective and efficient delivery options for their packages. The need was a solution to predict the estimated time of arrival (ETA) of delivery service providers at the retailer locations to pick up the packages and the estimated time of delivery (ETD) to end customers. The goal was to provide customers with accurate predictions and enable them to choose the most suitable service provider based on their delivery preferences.
SOLUTION The goal was to develop a Machine Learning model based on various input features such as distance, business traffic patterns, historical delivery data, and other relevant factors. The model would be integrated into the Delivery Solutions logistics platform to offer customers multiple delivery options with real-time ETA and ETD predictions.
We collected historical delivery data from their Snowflake data warehouse to build the training dataset. Our proprietary solution Zunō Predict, automatically selected the most relevant features from the dataset, built multiple models and picked the best-performing model. It uses a Signal Factory to generate and evaluate hundreds of features automatically, and then selects the best subset of features to train a model. We integrated the selected model into the client’s logistics platform, which allowed customers to input their package details and get real-time ETA and ETD predictions for multiple delivery service providers. Customers could choose the most suitable delivery option based on their preferences.
The model developed was successfully deployed, and the client saw significant improvements in their last mile delivery performance. The model’s accuracy was high which was a significant improvement compared to their previous ETA and ETD prediction methods.
By predicting ETA and ETD accurately, logistics platforms can improve delivery efficiency, reduce costs, and enhance customer satisfaction. With Zunō Predict, logistics platforms can automatically select the most relevant features and best- performing model, saving time and resources while delivering better results.
Integrating Zunō Predict’s Advanced Capabilities in Machine Learning and Analytics Will Help Us Gain Valuable Insights into Our Operations, Enabling Us to Make Better Data-informed Decisions and Boost Profitability.
Abbas Merchant,
Delivery Solutions Inc.
