Manufacturing executives have been overwhelmed with messaging around information and operational technology (IT/OT) convergence. That messaging suggests convergence must happen before they can take advantage of AI, machine learning, industrial internet of things, and other emerging technologies that will drive the factory of the future.
But from our discussions with mid-market manufacturers, “convergence” isn’t resonating–on the shop floor or in operations. It feels disconnected from reality and implies turning over control of operational technology to the IT function.
What manufacturers are looking for instead is productivity, efficiency, and agility. Operations engineers, process engineers, and regulatory personnel design controls to drive those goals. They need data insight into leading indicators to guide behavior and action. This often comes from IT because operational teams may not have the skills to generate and analyze data. But when data or analyses come back lacking appropriate context, operators don’t trust the data or follow the instructions. Thus, no real progress occurs.
The reality is that plant connectivity and operational equipment instrumentation have increased to such an extent in recent years that companies must do something about the gridlock in a time where VUCA (volatility, uncertainty, complexity and ambiguity) reigns for the foreseeable future.
We believe breaking this barrier requires a mindset shift in the conversation. Instead of beginning the discussion with AI, machine learning, or convergence, what if we focused on basic operational goals such as productivity, efficiency, reliability, safety, margins, or cost of goods sold? If we do that, how can we then use data to achieve that? And what do we need to do to make that happen?
Most data-related technologies have been the domain of IT and the back office. But as we know, operational technologies increasingly connect shop floor machines with sensors that emit data. Operational data typically has gone into the historian. Now, with cloud platforms, organizations can use big data to analyze operations, spot trends, and provide insight to the shop floor to improve productivity or address other goals.
Today, 75% of operational equipment is instrumented—creating a significant amount of data. What’s more, more than a quarter of manufacturing companies surveyed by the Manufacturing Leadership Council said their manufacturing data volumes had doubled or tripled in size over the previous two years. More than a quarter expect data volumes to surge by more than 500% over the next two years.
The success of future manufacturing operating models is dependent on collecting, managing, and using that data. We don’t see this as giving up control to IT; it’s about increasing data acumen and skills in operations so the two can work together more effectively to drive operational goals. When that happens, “convergence” occurs naturally—rather than feeling forced.
Several factors are driving the need for a thoughtful, comprehensive, and time-sensitive approach toward becoming a data-driven manufacturer.
With more connected IoT devices, there’s an increased threat of cyberattacks with a larger attack vector. Ransomware incidents are proliferating. Today’s risks of disruption, loss of intellectual property, and public reputation make this not just an operational concern but also a primary concern of management and the board. In our Q3 2021 executive poll, cyberattacks represented the second largest threat to conducting business, behind only hiring and retention of workforce.
Many manufacturers are struggling just to maintain enough of a workforce to run at capacity as they navigate the Great Resignation and other emerging workforce trends. The often mundane and repetitive work in factories can make it hard to retain people—even when paid well. But that’s just the beginning. Future work will be performed much differently than in the past or even today. Gartner predicts that 50% of factory work will be done remotely as soon as 2024. But while the factory of the future will be much more digitally enabled, it will still depend on people. Manufacturers cannot underestimate the significant amount of organizational change that will be required.
Any way to help extend the life of older equipment necessitates remote monitoring of critical machines, processes, and operations. This is particularly critical in an environment of high workforce turnover. Older operational technologies must be retrofit with sensor technology that can communicate with a cloud platform to enable reporting or data exploration. The hardware provisioning aspect of this alone is cumbersome and time-consuming.
Supply chain issues, pricing and margin pressure, political uncertainty, and the lingering impacts of the pandemic will continue to challenge operations for the foreseeable future. Data will be critical to making effective decisions in a VUCA environment.
Consumers have become accustomed to more choices: flavors, styles, colors, and customized elements. This means more changeovers in the factory. For example, where a plant used to make a brand of beer in several types of packaging, it now may make dozens of hard seltzer and energy drinks in an evolving array of flavors and cans. In this environment, production targets constantly change. Traditional manufacturing operations with quality checks or those that stick to the traditional recipe are not set up for such constant changeovers. This creates a drag on productivity.
This is a journey, not a project. It will take multiple years without a defined end point. Most manufacturers are just beginning their journey. While there is no one-size-fits-all approach, these concrete areas of focus apply in most situations. They can help organizations remain focused on the aim of using data to drive productivity, efficiency, and other operational goals.
Multi-year capital investment projects do not tie to annual P&L, and digital transformation or shifting to data-driven processes should be viewed as something that’s continuously evolving and not tied to an annual process. This lends a degree of added formality to initiatives, with specific goals and key performance indicators that are tied to capital management—a process the organization understands and knows how to manage.
Given the traditional divide between operations and IT, this type of change warrants a mission-specific team, empowered to define and take the steps necessary. If implemented in isolated internal “kingdoms”—as is too often the case—resistance will continue.
Resources should be dedicated to this mission to ensure a proper focus. If this is a portion of team members’ responsibilities, it becomes just another task for people who are already running at maximum capacity.
The team should at least include: operations/plant personnel and engineers as subject matter experts; IT experts who provide data and application support; and business unit (finance, sales, marketing) and distribution representatives to provide perspectives on enterprise needs, goals, and strategy.
Keep in mind that this is a general framework, and there is no one-size-fits-all formula. The team and structure must work for your organization. Also, leadership must take an active role in supporting by leading with a digital-first view, instilling governance with regular meetings to manage, learn from, and adapt the ongoing development.
The organization must look to quantify the return of becoming a data-driven manufacturer through identifying value with specific use cases, building the roadmap, and then capturing the value. For organizations with many locations, the roadmap should include standardizing a set of technologies specific to the vision. The complexity of legacy, siloed systems require considerable integration to enable a single source of truth throughout the entire organization.
Having data-driven operations will require new roles such as automation technicians, engineers, data scientists, IT/OT analysts, security experts, and IIoT solutions architects. Many of these roles require the skillset and professional competency of software engineering but with corresponding knowledge of operations and process engineering in areas such as quality assurance, maintenance, supply chain, and health and safety. That means you can’t just hire data scientists and drop them into your operations—and as a practical matter, that probably isn’t possible, given the market-wide shortage of data science skills.
What’s more, existing roles will change and require new skills. For example, operators need to understand how to interpret data presented through new interfaces, use that data to sense problems, and make decisions to prevent issues before they occur.
Together, these needs demand new and different approaches to recruiting, development, and retention than in the past. Training will need to become increasingly personalized—and manufacturers must be willing to pay for essential skills and offer incentives for reskilling.
Automation is critical to achieving productivity and efficiency goals. However, for most manufacturers, this isn’t about artificial intelligence, machine learning algorithms, and the “lights-out” factory—In a recent Manufacturing Leadership Council study, only 3% of respondents saw this as the eventual state of their factory model. It’s about providing more line of sight so operators can make informed decisions about difficult situations. Most manufacturers, then, will realize the greatest value by focusing first on the mundane: micro-stops and supervised intelligence tasks or digitized standard operating procedures. It’s also important to note future employees work better on screens than their counterparts with paper-based manuals.
To key to extracting maximum value from operations data? Sound processes for managing it. DataOps is the orchestration of people, processes, technology, and standards to deliver trusted, high-quality data that improves operational agility and speed of decision-making. All the different legacy systems will require data contextualization and normalization to enable that trust insight. When done well, it can drive significant productivity gains by getting information to the right individuals so they can increase control through a digital representation of physical operations. Think visible productive gains from standard operating procedures via an operator using a tablet on the shop floor. This delivers short-term returns while also gathering data for an algorithmic future.
There’s an important prerequisite for this: a clear picture of the data being generated and how it flows between and among assets and systems. The only way to capture this is to get in the trenches to identify, map, and document equipment/assets and data flow—and then put this information into a relevant management system. We see few manufacturers making this effort. Most understand the potential value but find it difficult to justify the cost or effort to do so. But without it, how do you expect to be able to connect, secure, enhance, and realize value from new technology–let alone the technology you already have? The longer you wait to begin the journey toward digital operations, the more complex and costly it becomes.
With the right foundation, you can do much more than optimize operations—you can begin to look at how to use manufacturing data as a strategic enterprise asset.
Experimentation with data technologies is still the norm in many manufacturing operations. Experiments allow people to see the actual impact—which can increase buy-in and drive momentum. But we see many organizations struggle to move beyond experimentation. The goal of experimentation should be to determine where to scale. If a use case has created value, you don’t want to lose that. Capture it and standardize its use across the enterprise to gain the greatest value.
The pandemic accelerated many existing trends and changed consumer expectations, buying habits, and lifestyles. We still don’t know what the new reality will look like–but that change will be part of our future.
Manufacturers need to be able to move quickly. Driving productivity or efficiency gains amid a VUCA environment requires the ability to use data and insight gleaned from sensors and connected systems. And that requires new skills and ways of working, with data competencies and practices resident within manufacturing operations. Think of this as a lifestyle change. The right improvements—better nutrition, exercise, sufficient sleep, mindfulness—can lead to sustainable results. As you make the key improvements, you'll begin to see the convergence that everyone is talking about happen naturally.