Integrating Supply Chain Data for Better Business Decision Making
Our client was facing several industry challenges including inventory forecasting, supply and demand management, transportation optimization, and a lack of data integration to make intelligent business decisions.
centralized data repository
order fulfillment due to reduced picking time
process efficiency and improved inventory turnaround
Background and Challenge Story
Our client, a leading global wholesale distributor of building materials, was facing several industry challenges including inventory forecasting, supply and demand management, transportation optimization, and a lack of data integration to make intelligent business decisions.
The client’s order fulfillment process required manual effort, which was time-consuming and prone to error. Not having the data to know the real demand for their products led to excessive production, which increased the cost of warehousing, transportation, and staffing.
They lacked integration across their supply chain data which prohibited timely and accurate forecasting and caused a considerable amount of product waste. They were manually loading their trucks, resulting in many trucks not being filled to capacity. Additionally, their warehouses were not organized efficiently based on product popularity.
Adastra’s solution needed to address the following: demand forecasting to leverage just-in-time inventory fulfillment, optimizing truckloads to reduce freight costs and drive sales margins, and optimizing warehouse layouts to reduce picking time and accelerate order fulfillment.
The overall goal of Adastra’s solution was to use data inputs that were readily available in disparate systems and tie them together. Then, using AI techniques, we provided value-driven indicators to either inform or influence business decisions around inventory, optimization, and customer experience.
In phase 1, Adastra implemented machine learning models to learn the patterns of buying behaviour and seasonality in the data. Adastra began with one customer for training purposes to help identify benchmarks, determine incremental impacts of new initiatives, and plan resources in response to expected demand. Following this, data from other customers will be ingested. Phase 1 allowed the client to implement a just-in-time inventory model, reducing product waste.
- Data Lake Storage on Azure Synapse and Event Hub
As part of the implementation, batch ingestion patterns were used to import CSV files into Azure Data Lake storage through Azure Event Hub and Synapse pipelines.
Next, real-time streaming was implemented through stream analytics and IoT sensor data.
- Continuous Integration and Continuous Deployment (CI/CD) Using Azure DevOps
The implementation of real-time streaming and sensor data collection provided the most up-to-date supply chain information, which was then automated through Synapse and DevOps pipelines, alleviating the need for repetitive manual effort by technical stakeholders.
- Extract, Transform, Load (ETL) Framework
Adastra worked to curate the data by cleansing and applying a transformation (ETL) framework in the staging layer, integration layer, and aggregated data warehouse layer in Azure Synapse.
In phase 2, Adastra leveraged the sales forecast insights from phase 1 to optimize truckloads, reducing the number of vehicles on the road as well as transportation costs. Multiple shipments can now be delivered in the fewest number of vehicles.
- Reporting Using Power BI
Using the output from the first two phases, a Power BI report was generated containing forecast summaries, truckload computations, and operational insights.
The client used the report to find recommended product quantities that needed to be prepared and shipped to specific stores, as well as truckload suggestions based on those recommended quantities. The order creation process can be automated in the future using a dashboard that will push the orders right into the system and get customer approval.
Phase 3 used the sales forecasting data and reporting from the first two phases. Knowing which items sell the fastest allowed the client to organize the warehouse efficiently and save time by placing popular items in easy-access locations.
The availability of data in a centralized repository enabled the business to generate valuable insights to make strategic business decisions.
- Informed business decisions
- Improved inventory turnaround
- Multi-million dollar savings
- Process efficiency
- Automatic data reconciliation framework
The client was able to reduce costs, due to optimization of truckloads and routing. Inventory turnaround improved with the ability to order adequately based on shelf life and demand and optimize warehouse layouts. Additionally, Adastra’s solution significantly reduced penalties paid for late deliveries. These efficiencies ultimately led to multimillion-dollar savings.
In collaboration with Microsoft, Adastra has been providing advisory services and thought leadership in Data Analytics and AI to the client since mid-2021.
Adastra is proud to have been recognized as the winner of Microsoft’s 2022 Data Platform Modernization, Financial Services, and Modern Marketing Impact Awards. These awards are evidence of Adastra’s ability to deliver value-driven data solutions to complex business challenges across sectors.