Manufacturers Collect a Lot of Data About Themselves—
and They’re Finally Starting to Use It
This may be the dawn of a ‘Second Machine Age’
Imagine you own a factory that makes smartphone cases. Business is booming, production is humming—and then, a key assembly line robot malfunctions. You contact the company that builds the robot. They promise to send a technician as soon as possible.
In the meantime, your line sits idle, costing you time and money.
But what if that technician could diagnose and fix the problem without ever setting foot in your factory? Better still, what if that robot—equipped with the right software and sensors, and plugged into an intelligent network—had let the technician’s company know two weeks earlier that it needed maintenance?
It sounds like science fiction. But actually, it’s the future of manufacturing. And that future is already happening.
Over the past decade, an information technology revolution has disrupted communications, media, and retail. For proof, look no further than the mobile device in your pocket, which can connect you to world leaders, stream live sports events—and help you buy a smartphone case.
Over the next decade, however, the ongoing upheaval in how we collect, analyze, and act on data figures to transform manufacturing, with increasing investment in new and improving technology remaking how we make things.
“Every 10 years, there’s a new cycle that tends to increase access to computing by 10-fold,” said Katy Huberty, head of North American technology hardware equity research at Morgan Stanley and co-author of a recent report on the convergence of information technology and manufacturing titled “Data Era Investment and the 'Second Machine Age.’”
“We’re at the tail end of the mobile Internet cycle, which was all about connecting consumers and creating and capturing the data that they create through engagement and purchasing and socializing. But 80 percent of the world’s data is actually locked within companies—and they’re just starting to analyze and get insights out of it.”
Manufacturing accounts for nearly 16 percent of total global gross domestic product, yet historically it has lagged behind other sectors in information technology spending. That’s changing. According to Morgan Stanley, business investment since 2016 within the United States on both information technology and industrial equipment has increased, the first time in decades that the two have risen in tandem.
Huberty expects that trend to continue and accelerate, with annual worldwide investment in information technology growing from $600 billion to $1.6 trillion by 2028. Roughly 40 percent of that increase is likely to come from industrial original equipment manufacturers (OEMs), per expert projections.
Why the rise? Improvements in data collection, processing, and storage technology have made manufacturing applications more affordable and attractive than ever before. Meanwhile, Huberty said, employees who grew up with access to the Internet now make up one-third of the workforce.
“We call them ‘digital children,’” she said. “When faced with a problem, their natural bias is to look to address it with technology. We’re now seeing that mindset in industries outside of tech.”
Factory floors have ample room for data-driven evolution and growth. To wit: since 2010, manufacturers have collected 2 million terabytes of potentially valuable data. Yet according to the Industrial Internet Consortium, companies have discarded 99 percent of that information without attempting to make use of it.
Robots are similarly underutilized. Morgan Stanley reports that American manufacturers currently average 84 robots per 10,000 employees—slightly above the global average of 74 robots, but far fewer than South Korea’s worldwide high of 631.
After coming into use in the 1970s, manufacturing robots mostly were mostly given potentially hazardous tasks, such as arc welding and painting in automobile assembly. For safety reasons, they were also largely caged off from human workers. That changed in the mid-2000s, when better sensors and software made robots less dangerous and more suitable for a wider variety of jobs—for example, working alongside people in electronics manufacturing.
Improvements in artificial intelligence (AI) and “machine vision”—that is, cameras and other sensors that evaluate image data and compare it with targeted outcomes, such as making sure an assembly line item is the correct color—are the next steps.
A Japanese company already has created robots that can teach themselves how to select objects from one container and place them into another. Meanwhile, a machine vision-equipped robotic arm that can select loose syringes from a table, fill and cap them precisely, and load them o to a distribution tray has helped a pharmaceutical material handling OEM double its production rate of filled syringes.
New technology also promises to improve old-fashioned human vision through assisted reality and augmented reality (AR), both of which superimpose digital images on what a user sees through special glasses or with a smartphone or tablet camera.
In Germany, a pharmaceutical packing machine maker is boosting its efficiency with the help of 3D goggles and a large screen that allows the company’s engineers to design equipment at real-world scale in a virtual environment. Quality-control workers at a Minnesota plant for agricultural equipment manufacturer AGCO are using AR glasses to speed up inspections and lower on-the-job training times for new assembly-work employees from 10 to three days.
“If there’s an issue on a unit being inspected, our workers are able to not only get information on that in real time, but take a picture or video of it and send it right back to experts who can help troubleshoot,” said Peggy Gulick, AGCO’s director of Digital Transformation, Global Manufacturing. “That’s a major advantage.”
Previously, Gulick said, employees would need to walk to and from nearby computer terminals to access inspection checklists and other information.
“We’ve saved up to 32 percent processing time on final factory inspections just by eliminating the waste of having to make that walk,” she said of AGCO’s AR system, which was implemented after 18 months of development and ergonomic study by the company. “And that’s the same on our assembly work—before wearables, our employees were walking to monitors to get the information they needed.”
According to the Morgan Stanley report, entire factories eventually will be transformed into intelligent hubs, with sensors and computers gathering and analyzing an increasing quantity of production and environmental data. In turn, this data will be used to make supply chains more traceable, foster greener and more sustainable practices, and more quickly identify inefficiencies and problems.
In Ohio, an auto plant outfitted with 259 robots, roughly 60,000 sensors, and a connected cloud computing system that keeps tabs on factory-wide wear and tear is saving 1.5 hours of production time per vehicle body while making as many as 830 a day. Elsewhere, Morgan Stanley reports, a paper mill using an advanced analytics system that allows engineers to remotely monitor production machines and send warnings and instructions to an on-site technician in real time has reduced maintenance costs by 50 percent.
Of course, there’s a potential downside to tomorrow’s data-driven manufacturing—high-tech solutions that eliminate jobs by replacing humans instead of augmenting them. At a Pennsylvania-based foundry that produces ceramic shells for the creation of everything from door handles to auto parts, for example, it takes six workers to make about 100 hand-dipped molds a day—but in another part of the facility, just three workers and a computer-controlled robot can make twice as many molds of higher quality.
Huberty said that the economic and labor policymakers and business leaders she’s spoken with believe that while data-driven automation may displace some manufacturing workers, it’s likely to cause less upheaval than previous technological disruptions such as the introduction of electricity and the automobile.
“On one hand, you have auto companies concerned that [self-driving cars] could put the four million people in the U.S. employed as professional drivers at risk—so how do you retrain them for new skills?” she said. “But in many cases, like with AI helping to read radiology charts, the number of radiologists has actually grown. And you’ve made them more effective at and satisfied with their jobs. So there’s generally not a view that this is going to be a devastation for human employment.”
Morgan Stanley projects that increased data and technology utilization within manufacturing ultimately may produce productivity gains equal to 5 percent of total costs. Consider a large energy company studied by Huberty. Previously, the company flew its senior engineers around the world to test and troubleshoot equipment—but now, those same engineers use AR glasses to work remotely.
“We’ve been starved for these types of technologies,” Huberty said. “When you talk to individual companies, there’s a sense that they see them not as one factor supporting their business, but rather what their business has to become over time.”
The author(s) and/or publication are neither employees of nor affiliated with Morgan Stanley & Co. LLC or its affiliates ("Morgan Stanley"). By providing this third party publication, we are not implying an affiliation, sponsorship, endorsement, approval, verification or monitoring by Morgan Stanley of any information contained in the publication. The opinions expressed by the authors are solely their own and do not necessarily reflect those of Morgan Stanley.
© 2018 Morgan Stanley & Co. LLC CRC 2315847 11/18