SRM Tech partnered with a specialized 3PL service provider serving retail and direct-to-consumer businesses to develop a sophisticated demand forecast engine that improves operational efficiency, eliminates spend leakage, and improves delivery SLAs.
Business Goals
The client is a third-party logistics company that assists multiple e-commerce businesses through order fulfillment, inventory distribution, and transportation services. Their internal process engineering teams rely heavily on forecast data provided by their clients to plan and execute warehouse operations. However, they have been struggling to arrive at optimal forecast estimates, which has had multiple adverse effects on operational performance as follows
- Throughput per Labor Hour (TPH) for fulfillment operations measuring labor productivity has decreased significantly.
- Increased instances of split shipments and higher frequency of DC-to-DC transfers leading to high transportation costs.
- Higher storage spending due to excess inventory holding.
- Plummeting customer experience ratings because of frequent shipment delays.
Solution
We partnered with their supply chain planning and process engineering group to analyze the existing data sources and forecast models. It was found that the forecast data supplied by the clients had an mean error rate of 55%. Furthermore, there did not exist any standard process for processing the original data from the clients.
Key Highlights
- Focusing more on data hygiene and decision quality, we analyzed and introduced improved and relevant parameters such as historical sales, sale prices, promos, discounts, product characteristics, market trends, seasons and weather patterns.
- We designed and trained an ML model with new and rich data sources and delivered a more accurate and scalable predictive modeling tool.
-
Leveraging this new demand forecast engine, the 3PL brand now has solid decision-making tools at its disposal and has scaled its daily operations to be more efficient and high performing.
Technologies Used
ML engine, Python, SQL, Streamlit
Outcomes
- Improvement in forecast accuracy by 37%.
- 12% reduction in order shipment-related SLAs.
- Throughout per Labor Hour (TPH) increase of 4% resulting in lower cost per unit (CPU).
Are you experiencing similar challenges in forecasting, inventory management, capacity planning, or labour productivity? Our expert team would be happy to extend assistance and deploy our solution accelerators to deliver value across your supply chain journey.
Drop in your enquiry here or email us at dpsales@srmtech.com!
