Skip to main content

Best Practice: Creating a Governance Mechanism for Analytical Learning in Organizations

I recently interviewed a large semiconductor company that developed a cross-enterprise initiative to drive analytical learning across its entire supply chain. The company established a supply chain IT organization that reports up to the CIO. It focuses on leveraging the corporate infrastructure and its SAP and SAP Hana implementation, in order to span all activities from order to cash and from demand planning to customer delivery and return. There are systems directors in both the procurement and logistics/planning teams. There are two IT support teams: one creates business process informatics centrally, and the other delivers locally required capabilities. The linkages between the enterprise and local teams are very tight, to ensure that functionally unique applications are inventoried centrally so all parties can leverage these tools. Recognizing the importance of having the right mix of data scientists and SMEs, the company has established a supply chain information and analytics organization composed of data scientists, operations research, and statistics PhDs to work on problems they are directed to from the business.

An important governance mechanism that drives analytical process is the Supply Chain Director’s Council, whose mandate is to focus on the availability and ability of the supply chain to deliver parts and core finished goods to customers. All of these groups (supply chain IT, systems, supply chain analytics, planning and logistics analytics) participate in the Supply Chain Director’s Council. Because of the long lead-times associated with the manufacturing infrastructure required to source and produce semiconductors, the Director’s Council espouses a long-term vision and tactical planning scenario. They work with the Supply Network Steering Committee (senior executives in supply chain and IT) who set the direction for the company’s supply chain capability. This includes setting direction for how the company will use emerging technologies (e.g., 3D printing, IoT technologies, and artificial intelligence). However, the Director’s Council defines the critical needs and availability requirements for existing and mature product lines, as well as experimental and emerging growth categories. Questions addressed include, “What will be the next generation of products?” The team also leverages insights from university research centers and consulting groups, including CAPS Research, MIT, XEM, Gartner, ProcureCon, and Procurement Leaders, to ensure they are connected to emerging ideas. They are also involved in pilot programs with multiple customers exploring the application of IoT and analytics approaches, with customers such as Caterpillar, P&G, Monsanto, Amazon, and Unilever. These pilots are used to provide input into how the enterprise wants to design data collection and analysis structures; for instance, how environmental requirements impact wafer fabrication, constrained timeframes, and temperature. The supply chain analytics team is involved in a great number of simulations to address these industry-specific local questions. The team acknowledges that there is much work to do in predictive analytics, to create mature systems for on-going operations, and to be able to collect insights from third parties.

As an example, the company is beginning to triangulate analytical approaches with external data to drive improved planning. Demand planning is influenced by GDP growth factors, how this affects the movement in semiconductors, customer refresh cycles, current order status, and seasonal variations. It also considers leading indicators that provide clues as to what customers are planning to ship, interesting correlations that have been developed, and hypotheses and their implications for internal capital investment and capacity planning. This is broken down into more minute analyses of what could happen with a smaller range of capital, and what could be delivered with different product health indicators including yield or other operational metrics. All of these pieces get meshed together and reviewed by a pricing team and corporate economists who understand the macro trends and impact of pricing changes on the market. There is a fine line between investing too little in capacity that would leave money on the table, and investing too much in capacity that would incur a waste of investment dollars. The company always tries to err on the side of supporting customers at the highest level.

The data analysis today is intense, involving a huge amount of data crunched for review. According to the team, “One of the biggest limiters for us is the ability to ask the right questions!” The team relies on the aggregated insights of practitioners and analysts in solving problems. Practitioners bring the operational reality of construction programs to light so the environment can be modeled correctly, while analysts provide skills like Monte Carlo simulation that can model the range of variability and the level of risk each scenario implies. The outcome of these simulations is used to plan supplier capacity and supply network design, with mix and model monitoring information passed on. The company recognizes that both mobile and cloud computing are new vectors that must be planned for, and the new technology group considers these developments in its long-range strategic plan. Leaders believe that two of the biggest changes for supply chain will be in robotic automation and the application of artificial intelligence and how we improve the learning of the workforce to leverage the tools and available data to parse and leverage talent in new ways.