Executive Summary
One of the largest shipping companies in the world, that offers land, air, and ocean logistic services, wanted to improve their ocean trade logistics. The company experienced a significant spike in demand post pandemic and wanted to streamline operations. It was a challenging task as they operated at multiple trade levels (specific geographic region or trade route) and across multiple trade corridors.
The client desired to enhance the current inventory management dashboard, which had a prediction accuracy of 40-60% for demand. Their objective was to assist customers in better optimizing their inventory management for seamless operations. The team at Factspan used historical data available to build a new machine learning model to boost prediction accuracy and provide valuable insights for proactive inventory optimization strategies.
About the Client
The client is one of the largest shipping companies in the world offering ocean and inland freight transportation and associated services, such as supply chain management and port operation.
The company is based in Europe, with subsidiaries and offices across 130 countries and around 80,000 employees worldwide in 2020. They serve various sectors that include Fast-Moving Consumer Goods (FMCG), retail, chemicals, fashion, and lifestyle.
Business Challenge
Due to the company’s extensive trade operations across multiple trade levels and corridors, the port authorities had to handle a large inventory to meet customer demand. However, they faced the challenge of accurately predicting the volume of containers required at the port, leading to a shortage or excess of inventory space at ports. The scenario had the potential to cause customer dissatisfaction and impact the company’s revenues.
The customers, who are involved in shipping assignments, needed to determine the storage capacity at the port to ensure they could accommodate enough containers to meet the demand for their goods. The organization wanted to improve accuracy and the efficiency of the forecasts to ensure the customers will not lose out on any business due to inventory management challenges.
Our Solution
Factspan addressed the client’s business challenge by gathering historical data on the trade logistics demand. They used a traditional modeling approach with Causal AI to develop a demand model that accurately predicts trade levels for different trades and corridors.
To ensure inventory optimization, Factspan enhanced the existing dashboard for users, by focusing on bridging the gap between forecast data and actual information. The dashboard now shows the predicted demand for a week / month, allowing the customers to manage their resources accordingly. It also allows them to plan their operations better and reduce the time and cost required to manage inventory.
Factspan’s solution improved demand prediction accuracy and maximized throughput, leading to increased customer satisfaction and improved profitability for the shipping company.
The model was built on the following algorithms:
Pearson correlation coefficient, RFE with LGBM, Random Forest – SelectFromModel, LGBM Regression – SelectFromModel, XGBoost, Catboost, RFE with Lasso, Spearman correlation coefficient.
Business Impact
- Improved demand prediction accuracy from 40-60% to 80%+
- Optimized inventory management for client’s customers boosting output by 20-40%
- 70% increased customer satisfaction by ensuring the availability of inventory
- Improved profitability for the client by reducing inventory costs by up to 25%
Most Popular
Shipping Storage Risk Prediction using M...
Gen AI-Infused Business Analytics for En...
Boosting Efficiency with Gen AI through ...
