Navigating Automotive Disruption: Eliminate Bottlenecks in your Supply Chain and Lean-out Operations to free-up Working Capital for R&D

The Automotive Industry is being disrupted on all sides and is upending the traditional models of vehicle manufacturing that have evolved over the past 120 years to dominate the world. New market entrants like Tesla have reinvented the electric power-train, created their own energy ‘fuel’ distribution grids to end-point superchargers, and redefined the once ‘ludicrous’ notion of what an instant-torque, Blitz-acceleration, luxury  status-symbol with actual ‘Fahrvergnügen’ can be.

Development cycles for new models have also been compressed by aspiring entrants, both new and foreign, while with Over-The-Air (OTA) software updates now download new features, functionality, and refined interfaces to existing cars and trucks regularly and without a trip to the dealer, nor a new model upgrade.

As cars have evolved into rolling networked supercomputers, security remains a top issue, but has been giving way steadily to the convenience and customization that modern buyers are increasingly demanding in the connected world of mobility. Google’s Android and Apple’s iOS operating systems are pressuring OEM’s to open-up their previous walled gardens around clunky Entertainment, Navigation and Apps to provide more open, intuitive and personalized experiences.

Most recently, Autonomous Vehicle Technologies have accelerated their penetration into mass consumer vehicles, where the ever-more sophisticated merger of new hardware,  software and sensors are bringing about new advances in convenience and safety. While the safety statistics generally support this trend towards increased autonomy, the fringe cases of machine-learning and (semi)-autonomous navigation gone awry, however, still dominate the news cycles. Meanwhile end-consumers may be turned into unwitting beta-users for real-world field testing and test data generation before such new (semi)-autonomous upgrades are fully baked.

And connected cars now provide not only in-vehicle WiFi-over-Cellular Internet hubs for families, but also make new ride-sharing, on-demand usage, and vehicle-ownership business models possible. Connected vehicles furthermore enable the collection, analysis and modeling of massive data sets, including vehicle, services and fleet related. Subsets of such data, and its derived meta-data, is now already being back-hauled across networks and joined with many other proprietary data sets, including decades of aggregated consumer and sales data. Further joined with other data sets, including manufacturing data, maintenance data, and recall data, such connected “Internet of Things” (IoT) sensor data generated from many vehicles individually can enable better operations, planning, and predictive maintenance, across entire models, makes, fleets and geographies. Even Ford CEO Jim Hackett talks about auto manufacturers’ decades of proprietary consumer, vehicle and test data as a market differentiator. Combined with increasingly more connected vehicle and IoT data sets, he sees Ford’s unique data advantage and huge “monetizing opportunity versus an upstart”, with their smaller fleets sizes and only limited new data sets to work with across just emerging customer bases. Of course, data may be “the new oil”, but if not refined properly through analysis, and not put into the right tank for action by domain experts, then it’s never going to propel the motors of corporate innovation forward. After all, rocket fuel in a diesel motor won’t get you far on land or in space.

Yet industrial-scale disruption on multiple fronts requires lots of innovation. And such innovation requires new ways of thinking, different methodologies, and R&D - not just heaps of it, but mountains  of it. Whether making the decision to “build, buy, or buddy”, that is, through internal R&D, acquisition of technology / companies, or licensing external technologies through partnerships, respectively, adequate capital is still a key determinant factor of success or failure. Whether you’re an Automotive incumbent still producing traditional vehicles, or a new market entrant flush with venture capital or deep-pocketed investors, how does one free up the additional working capital required for such new R&D investments to begin with in an otherwise entrenched, hyper-competitive industry? After all, automotive operations by their very nature are large-scale, cross-border, just-in-time, factory-floor endeavors of gigantic proportions with global supply chains that aren’t easily reproduced.

Even the ultimate modern disrupter, Elon Musk, has been humbled time and again by the mundane world of operations, while missing virtually every output target in trying to scale-up beyond smaller vehicle production volumes towards a global mass-market vehicle manufacturer. With great ambition comes many failures along the way, and without highly efficient operations conserving cash, further capital is required to fund product sprints, fast failure, accelerated learning and growing pains. Mass-scale global manufacturing is hard after all, as today’s leading incumbents have learned the difficult way over the past 120+ years, even while shouldering the inefficiencies of contemporary automotive manufacturing and markets.    

But the Automotive industry's massive scale can also be its savior in an age of unprecedented disruption across multiple fronts. As Shingeo Shingo, a pioneer of Lean Manufacturing, who also formalized the Toyota Product System (TPS) process, said,  “The most dangerous kind of waste is the waste we do not recognize.” More often than not, that is the waste hidden in plain sight right in front of you and around you. It’s the status quo and embedded part of the process simply accepted ‘as-is’, and not peeled-back in more detail to understand each step within the process and where additional fat can be trimmed out of the system. Such waste is endemic in every operation and every process of every kind around the world. And in the world of Automotive, with end-to-end supply chains that span the globe, large factories churning out heavy vehicles, hundreds of Tier-1, Tier-2 and other suppliers with their own manufacturing, supply chain and operation, and multiple shifts of many types of workers of all skill-sets frequently interoperating around automated and robotic systems, there are untold, unseen systematic wastes of all kinds at every level that can be leaned-out of the processes, both large and small.

In fact, the industrial automotive leaders today each harbor tens-of-billions of dollars of annual waste per manufacturer, where working capital can be unlocked and reinvested in innovation, R&D, new types of worker training, modernized manufacturing lines, more efficient safer systems, leaner processes, greater profitability, higher Earnings-Per-Share, and ultimately more investor distributions and share buy-backs.

In short, global automotive manufacturers are leaving hundreds-of-billions of dollars in aggregate on the table by not embracing a more data-driven culture that eliminates waste and empowers their existing domain experts to achieve more with less for better processes and end-results. Today it’s their consumers, employees and investors who are paying the price, and some would argue the communities, societies and the environment they live in and around as well. After all, wasteful production meets no meaningful need, yet consumes resources better utilized, or conserved, elsewhere.

But to achieve superior results, both traditional and new methodologies must be taken into account when approaching engrained issues afresh. For example, certain business concepts have been around so long and are captured in such well-known, catch-all phrases, like the ever-present, all-encompassing “Supply Chain” or “Six Sigma”, that they seem to lose their relevance in everyday business, such as automotive. Nowhere is this more apparent than in Silicon Valley, where the constant influx of “innovative” and “disruptive” technology entrepreneurs seem hell-bent on upending entire industries and disintermediating market share for themselves, without necessarily comprehending the massive failure rate and the ability of some incumbents to still innovate and leverage their considerable industry girth, especially when attacked on all fronts.

In fact, every product, good or raw material is supported by a supply chain, and in many cases, such as automotive or electronics, those end-to-end supply chains are global in nature and components, sub-systems, and systems can pass over dozens, even hundreds, of borders before a consumer purchases the final product. As product companies leverage greater data and try and expand into new services, those too are kept afloat by invisible supply chains and support networks spanning the globe, with each segment and system generating data, both old and new archetypes, that can be leveraged for a greater corporate good.

And for over half a century, these supply chains, from simple raw materials to complex automotive products with tens-of-thousands of components, have been running in part thanks to software, hardware and sensors that produce data. Not the more recent hype of Big Data, Industrial IOT or Industry 4.0 time-series data, but basic data sets - smaller, siloed, disconnected, and then often forgotten to almost anyone outside the realm of Operation Technology (OT) and Operations on the factory floor and in the control tower. These Smaller Data sets, however, are live in production and a forgotten gold-mine of bountiful time-stamped data. High-tech Industrial IOT edge sensors may generate petabytes of data from jet engines and connected fleets, but the precise and dependable time-stamped records in many key legacy systems actually show what’s going on across your supply chain at that location across multiple points in time. If and only if you know how to tap-into and leverage these multiple, disparate, disaggregated legacy data sets. To the point, they’re often captured in simple relational databases, log files, historians, data system files, legacy systems, even spreadsheets and the like. These data sets, stitched together over time, prepared, standardized, synced and properly analyzed from the point of view of multiple best-practices and heuristics, all in parallel, can provide a real-time view of the health of your operations. This can include  invaluable insights on which bottlenecks are building-up and occurring now, what steps you can take to eliminate them, and how much you can save, or lose, by addressing them expeditiously, or not at all, respectively. A so-called Bottleneck Management System (BMS).

These Small Data sets don’t have trendy names like 5G and Industry 4.0, but rather, the tried and true weathered acronyms few outside of OT understand: PLC (Programmable Logic Controllers); MES (Manufacturing Execution Systems); IMS (Inventory Management Systems); TMS (Transportation Management Systems); WMS (Warehouse Management Systems); EAM (Enterprise Asset Systems); SCADA (System Control And Data Acquisition); DSC (Distributed Control System); MRO (Maintenance Repair and Operations); Historian; Field Buses, and yes, even the more common but no less important, ERP (Enterprise Resource Planning), CRM (Customer Relationship Management), and POS (Point Of Sale), as well as simple log files, .CSV, .TSV, ASCII and many other data file types far too numerous to list here.

But if you utilize a Bottleneck Management System, which combines these disparate data sources, utilizes advanced signal processing, synchronizes their frequencies, automatically prepares the data, and processes such diverse data sets automatically and in parallel using Machine Learning and Artificial Intelligence (AI) techniques, both historical for long-term strategy planning, and where available streaming, for improved real-time execution, then you can start to get a much more nuanced view of your entire supply chain. For example, where are your operations most efficient? Where are they not? What bottlenecks are holding up specific processes at any given point in time? Where are bottlenecks are building up and why? How do you reassign resources to address? Where can you free up more working capital from inventory? How do you increase your cycle times and inventory turns?  How do you best address them based on past analysis and real-time streaming data? What are the highest-ranked prescribed solutions for the most similar issues? How much you can save, or lose, by addressing such hindrances, or not? Which of your domain experts and resources are best suited to address the problem within the given timeframe and with a preferred, most likely outcome?

Mr. Takao Sakai, a leading global Lean Manufacturing & Just-In-Time expert to the largest conglomerates in the world, harks from the family of Dr. Kotaro Honda, the father of the applied "Just-In-Time" philosophy, inventor of Kichiei Sumitomo (KS) Steel, Nobel Prize-nominated Physicist, and investor who helped Kiichiro Toyoda with his startup ‘reboot’ after WWII, with the Toyota Motor Company pivot. Mr. Sakai points out that, “Through software, we can now democratize Just-In-Time to be available at unprecedented scale to all Kaizen professionals and related areas such as Lean Manufacturing and Lean Six Sigma. With Industry 4.0 and IOT initiatives around software, there is excellent potential for manufacturers to optimize end-to-end production processes throughout the product life-cycle”.

Yet identifying and accessing all this OT data first across diverse legacy data systems, and then processing the most relevant sources in parallel with such software to solve your business problems, however, is not a simple proposition, and is well above the ability of even the most exceptional human calculators. This is where Machine Learning and AI comes in: it’s not the end-all, be-all that the high-tech industry is trying to sell-you, but rather, simply a tool in your arsenal to provide a more efficient means to a very desirable outcome when implemented correctly. That end goal being better results across your entire operation that meet your everyday operational goals, improve your KPI’s (Key Performance Indicators), translate directly down to the bottomline, and ultimately increase the Earnings-Per-Share (EPS) estimates that the C-suite, Board and investors are focused on.

And those results can be distilled down into five main pillars of lean automotive operations, KPI families and ultimately desirable beneficial outcomes, namely:

1) Leaner, 2) Better, 3) Faster, 4) Smoother, 5) Safer

Or said otherwise from an Operations Perspective, mapped alternatively as:

1) Inventory, 2) Defect Rate, 3) Cycle time, 4) Uptime, 5) Changeover

If you’re looking to increase Operational Efficiency no matter where you are in the world, what your business is, and what type of constraints you have, you only have so many variables to play with. And in automotive, with worldwide operations across borders with many types of constraints, including labor, tariffs, safety, environmental, fuel-efficiency, emissions, quotas, part availability, supplier complexity, shipping, strikes, recalls, and many more, it is all the more complex. The question then is: when you have so many moving parts and variable processes across your supply chain, particularly with the inherent complexities of Automotive, and have so many disparate legacy data sources, how do you keep on top of it all with a limited team, dated tools, dirty disaggregated  data, static BI analytics software, an aging workforce, budget pressures, and market disruptions from all sides?

First off, focus on small repeatable projects that prove out value using focused operations-specific software on the data you already have, and leveraging the domain experts that already exist in house, to solve the problems, seen or unseen, that have plagued your business supply chain for years. You need not tap all your data sources at first, just a few will do to prove out value and demonstrate a valuable Return-On-Assets (ROA) or Return-on-Investment (ROI).

Second, realize that only 4% of organizations are even attempting to use AI in their Supply Chains for a reason. Instead of trying to hire a small army of expensive Data Scientists to manually crunch your dated data with algorithms after the fact, instead focus on using straight-forward, operations-specific software solutions. Bottleneck Management Systems must automatically incorporate AI and Machine Learning transparently and seamless in the background, so that your everyday operational teams can use the software dashboard as simply as a spreadsheet or business intelligence tool on both your real-time and historical data sets.

Third, lean-out and get your own house in order first, adding additional data sets as you go along to flag more and more bottlenecks that are constricting the throughput, output and profitability of your operations. Then once you have a smooth flow, you can focus more externally and squeeze out better terms and concessions from suppliers, contractors, service providers, and other marketplace players. After all, there is waste right there in front of you, hidden in your Operations and Supply Chain. Yet it lies within your power to identity, quantify and control. Thus allowing you to unlock the savings and efficiencies that can be gleaned between the lines of your own data with the enablement of your own domain experts and some help from automated, scalable AI BMS software that’s easy to use. It’s the low hanging fruit. Pick it first before your competitors do and reap the bounty while it’s still ripe for the picking and not rotting in your warehouses and showrooms as expired inventory.

Fourth, resist the urge to go after those massive ‘Digital Transformations’ with industry experts who don’t really know your company and industry domain like you and your own teams already do. The rule of thumb is that if you can’t find savings and fix the low-hanging fruit of your own business, then why do you expect outsiders who know less about your business to do it better? You’re going to pay them to distract your own people and teach them the intricacies of your own business, and then pay them even more to make recommendations, many of which are already second nature to you and your teams. Ask yourself why. If your unique domain expertise isn’t yet automated, institutionalized or even properly recorded, then use operations BMS software with AI that incorporates Continuous Improvement, Kaizen AI, and Yokoten feedback within the software offering. After all, as your workforce continues to grey and goes into retirement, all that institutionalized knowledge and best practices sitting in their collective noggins walks out the door with them. Replacement training is helpful but too much is lost in translation and timing. A good software package helps facilitate the capture of institutionalized knowledge within the operations module themselves, allows annotation and sharing across the organization, suggests and prescribes the most likely fixes, to support your teams of domain experts in making their own calls based on the unique situation at hand, and similar situations that may have occurred in the past, across a  multitude of same or similar facilities. AI BMS software can help streamline, suggest and enable, but your people still make the ultimate call and take the best call to action.

So instead of outsourcing the responsibilities to others, step up with a series of one or more smaller engagements with a few data sets, and prove value at a smaller scale with your existing smaller teams sharing their knowledge via software at one or more facilities. Kick-off these projects within your current budgets and prove-out the shorter-term wins, learn from the bumps experienced, and then scale-out to larger projects with more data across more facilities. As you achieve more success, you’ll realize more savings, thus freeing-up more budget organically across more locations. In this manner, you can better continue to aggregate and layer-in more data for analysis to flag ever-more of the always-shifting operational bottlenecks that are holding you back in your own supply chain optimization, thus unlocking more working-capital with every project expansion and further fueling the efficiency gains. It’s a positive flywheel effect that multiplies over time and can leave your competitors in the dust.

Few industries have supply chains as global, complex and inter-dependent on many Tier 1 and 2 suppliers from countries around the world as automotive. So while the automotive industry is being disrupted and attacked on all sides, it would behoove automotive operational execs and leaders to avoid their fiduciary responsibility and not pursue projects that can prove-out and bear near-term, quantifiable benefits with savings in the millions to billions of dollars annually at scale.

If you’d like to learn more about Bottleneck Management Systems leveraging AI on your existing data to empower your existing teams already today, send an email to set up a call at: [email protected]

I look forward to your insights, perspectives, and discussion.

Seth Page

COO of ThroughPut, Inc.

P.S. - For deeper, unfiltered insights on how to improve your operations with your own data, your own teams, and AI-enabled BMS software solution partners available today, please refer to ThroughPut’s CEO blog on:  “The 7 Industrial Data Myths Preventing Your Operations From More Data-Driven Profitability and Productivity” at:  

https://www.dhirubhai.net/pulse/7-industrial-data-myths-preventing-your-operations-from-raza/