One of the major challenges in managing supply chains is coordination and communication within the different stages. We encountered this issue at Slipintosoft, where our supply chain was suffering from inefficiencies leading to delayed deliveries. To rectify this, we turned to data analytics, identifying key areas of bottleneck and delay through data visualization techniques. For instance, we noticed raw silk inventories were piling up unnecessarily at our factories due to a lack of real-time data on demand, substantially affecting our cash flow and storage costs. Elasticsearch and Kibana became substantial tools in bridging this information gap. With real-time insights into consumer demands and product lifecycle, we were able to streamline our inventory, reducing unnecessary holding costs by about 30% and increasing overall supply chain efficiency. This incident has solidified my belief in the transformative power of data analytics in supply chain management, extending its benefits beyond identifying problems, all the way to increasing operational efficiency and ensuring customer satisfaction.
An efficient and robust supply chain for retail operations involves various complex decisions that depend on each other. Viewing the supply chain comprehensively requires integrating large volumes of data from multiple sources (sales, suppliers, market trends, etc.) and implementing real-time monitoring systems to track inventory, supplier performance, and transportation to enable quick responses to any issues. Big data helps with these decisions by identifying potential risks and optimizing supply-chain operations. A recent example of using data-driven insights is when COVID-19 significantly changed the last-mile delivery network for B2B retailers. Due to work-from-home policies, delivery locations were more spread out in the suburbs, with one to two boxes per stop, compared to a larger number of boxes in more centralized office locations. This exposed a general underlying inefficiency in last-mile delivery when (total route time) demand exceeds (temporal) delivery capacity, where the challenge is to deliver to all customers on the promised delivery day with the retailer-driver (RD) staying within the regular shift hours. Evolving industry practices include outsourcing some deliveries to on-demand drivers (ODDs), such as Uber and Lyft. I used big data and AI models to help determine the route of the RD, the locations that the ODDs will deliver to, and the drop-off locations where the RD will hand over packages to the ODDs.
Our client, a leading industrial supply company in Southeast Asia, faced the challenge of managing vast amounts of inventory-related data. To address this, they sought to extend their existing data warehouse solution and reduce operational costs associated with on-site consultants. Additionally, the company aimed to migrate the solution to the cloud for enhanced scalability and cost-efficiency. To achieve these objectives, we developed a unified cloud-based big data platform capable of handling large datasets and automating data extraction processes. The platform integrated over 100 data sources, including daily data loads and historical backfill, and successfully processed terabytes of data. One of the key challenges was migrating from on-premise platforms to AWS. With this in mind, we built a new AWS-based big data platform from scratch and extended the existing data collection and reporting functionalities. We also integrated with various data sources, including MS SQL, Oracle, and SAP. Next, we conducted a proof of concept to select the optimal data warehouse design and technology stack, comparing various solutions to ensure cloud neutrality and the ability to scale resources flexibly across different cloud providers. The entire development process was designed to be cloud-agnostic, allowing for easy future transitions between cloud providers. Terraform, a tool compatible with AWS, Azure, and Google Cloud, was used to ensure flexibility and portability. Through this project, our client successfully transformed their inventory management capabilities by leveraging a powerful cloud-based big data platform. The solution enabled efficient data management, reduced operational costs, and provided a scalable and flexible foundation for future growth.