Applied to the automotive supply chain, predictive technology can be used to spot real-time trends and patterns in supplier pricing and performance data.
Getty Images
If you had visited General Motors’ Oshawa plant in southern Ontario this past summer, you would have encountered a stark yet telling image: over 95,000 unfinished vehicles sitting in storage lots awaiting completion – a vast monument to the impact of the current supply-chain crisis.
In recent months, supply constraints have forced automakers such as GM to adopt strategies such as “build shy” – building vehicles to completion shy of just one or a few parts and storing them until they can be completed. While many consider “build shy” a better choice than starting and stopping production lines, it can also create substantial risk for automakers in the form of huge vehicle backlogs, storage costs and unrealized revenues (not to mention potential depreciation).
With inventories still inflated from the summer and an economic recession appearing ever-more imminent, now is the ideal time for automotive OEMs to reflect on the decisions that got them here and exploit some of the innovations that have arisen due to parts shortages, especially microchips.
The Predictive Revolution Comes to Automotive
Applied to the automotive supply chain, predictive technology can be used to spot real-time trends and patterns in supplier pricing and performance data. Using a predictive model, procurement teams can even preempt a supplier’s quote – effectively naming the price they are most likely to accept and communicating this price and terms to multiple suppliers. This practice of communicating the desired outcome for a purchasing or sourcing cycle to suppliers offers a window into the future of a self-driving supply chain – one where certain parts and components more or less procure themselves.
However, no matter how well automakers plan, sheer unpredictability can throw even the most well-oiled procurement and supply-chain operations off course, creating bottlenecks in needed parts and supplies. The current chip shortage is a case in point. When COVID-19 hit, demand for new automobiles plummeted, and chipmakers pivoted their allocation toward the high-tech sector to maximize their profitability. When vaccines became available and automotive demand picked back up, resource-constrained procurement teams couldn’t move fast enough, and found themselves at “the back of the line” with many chipmakers who prioritized larger orders with preferential payment terms.
This scenario demonstrates the need for greater procurement agility in the face of constant and accelerating uncertainty, and greater agility demands increased speed (e.g., shorter procurement cycles, faster approvals and the ability to adapt administrative processes to changes in the market). Predictive procurement helps automotive manufacturers pre-approve a price and thus cut steps (like asking for a quote) out of the process completely, thus putting procurement in the driver’s seat and taking them out of a reactive posture with ever-increasing changes in the market. The key element underlying the predictive revolution in automotive is continuous access to real-time market trends, including an audit trail of the automaker’s past price agreements with a cohort of relevant similar suppliers.
Understanding the Evolution from Should-Be Costing to Predictive Procurement
In most industries, cost is the key factor when selecting suppliers. One aspect of automotive procurement that makes it so complex is the need to consider volatile costs alongside other quantitative operational dimensions such as lead time, minimum order quantity and shipping terms. Traditionally, OEMs have dealt with this complexity using a modeling technique called “should-costing” or “should-be costing” that creates a bottom-up cost model for an individual part’s components (including labor and shipping) and then uses this model to check pricing across a complex market basket representing a vehicle’s Bill of Materials.
The challenge in using “should-be costing” in a disrupted market is that volatility will impact every supplier differently, meaning that price offers are likely to diverge significantly depending on source of supply (a phenomenon known as “price dispersion”). This “one-size-fits-all” pitfall of cost data in automotive procurement requires a new approach that persistently simulates, predicts and mass-personalizes cost models across an entire supply base.
Predictive Technology Will Enable Self-Driving Automotive Supply Chains Well Before Self-Driving Cars
Predictive technology is poised to revolutionize the automotive industry in many exciting ways. However, predictive procurement may offer an even bigger bottom-line opportunity for automakers and their suppliers in the short and medium term. By giving suppliers faster purchase orders and procurement teams confidence in their approval decisions, predictive technology can create “win-wins” for supplier relationships.
Quote analysis and benchmarking is just one example of how predictive procurement is being used by large automotive OEMs to solve real-world problems with tangible impact, ultimately improving operations, maximizing working capital and driving greater revenues and profitability.
Edmund Zagorin (pictured, above left) is founder and CEO of Arkestro, which leverages emerging technologies to shorten and automate the procurement process. Eric Buras (pictured, left) is head of machine learning at Arkestro.
Keyword: What The Predictive Revolution Means for Automotive Supply Chains