Back

March 1, 2014

Big Data’s Power for Manufacturing

Why you need to care about the digital revolution

The trucks and pump trailers for hydraulic fracturing that Texas-based M.G. Bryan Equipment Co. sells and leases can cost as much as $1 million each. These vehicles operate in extreme, isolated environments and are expensive to operate and repair. They require new oil filters every 200 to 400 hours and a total engine rebuild after 4,000 to 7,000 hours of operation. Downtime can cost as much as $7,000 a day.

In an effort to reduce those costs and maximize the ROI on this equipment, the privately held, medium-sized company worked with Rockwell Automation, an industrial automation software and control systems developer based in Milwaukee, to create a system that could remotely access real-time information, and produce automated maintenance alerts, and parts and service delivery requests. To accomplish this, Rockwell Automation used state-of-the-art remote equipment sensors, data collection and analytical software to monitor each individual vehicle’s condition and fracking performance, as well as provide maintenance trends for an entire fleet.

The result: increased uptime, improved productivity and, in one case, the ability to respond within minutes to a pump-engine surge that could have burst a pipe and caused worker injury, a 100-mile service trip and $60,000 in replacement parts.

“We see [right now] as an inflection point for IT systems and production systems to come together,” says Sujeet Chand, chief technology officer for Rockwell Automation, the world’s largest provider of industrial automation. Chand notes that manufacturing generates a lot of data, but most of it is still not collected, analyzed and managed effectively. 

Metals companies in particular have been relatively slow to adopt these new digital industrial capabilities. As with so many companies, their computer systems may not be well integrated: The sales system might not communicate with the inventory and production systems, for example. Little, if any, of the equipment or materials they manage may be remotely monitored. Management simply may not see the value in spending often-significant amounts of money on new systems and new ways of doing business. But increasingly, for businesses large or small, those who fail to fully explore the potential for increased efficiencies and profitability from sophisticated data collection and analysis will find their market position and even their survival facing intense pressure.

Some people talk about the Internet of Things or the Industrial Internet in an effort to define the expanding role that Big Data and data analytics play in changing the way a growing number of companies do business. This is the world shaped by an increasing volume, variety and velocity of data—where, more and more, machines communicate with machines, and humans tap in to monitor results or rethink operational insights and changes based on data mixes and complex algorithms.

The economic potential of Big Data for manufacturing alone is massive—as much as $270 billion annually in increased GDP by 2020, according to a 2013 report from the McKinsey Global Institute on “game changers.” Half of that increase for manufacturers, the report said, could come from data-driven predictive maintenance and analytics.

 

The Second Digital Economy 

This change is what economist and technology thinker W. Brian Arthur calls the Second Economy, comprised of “all of these digitized business processes conversing, executing and triggering further actions.” This second, digital economy, he wrote in the McKinsey Quarterly, is “vast, interconnected and extraordinarily productive,” capable of generating massive economic growth and representing the most significant shift since the Industrial Revolution. It is “helping architects design buildings, it’s tracking sales and inventory, getting goods from here to there, executing trades and banking operations, controlling manufacturing equipment, making design calculations, billing clients, navigating aircraft, helping diagnose patients and guiding laparoscopic surgeries.”

A 2013 Gartner Inc. survey of 720 businesses from diverse industries found that 64% are either investing or planning to invest in Big Data initiatives, up from 58% a year earlier. Frank Buytendijk, Gartner research vice president, characterized 2013 as a year of “experimentation and early deployment,” with process efficiency representing the top priority of manufacturers.

 

Understanding New Value Potential

A.M. Castle & Co., an Illinois-based metals processor and supplier that operates in more than 50 locations worldwide, is still working to capture the full potential of an enterprise resource planning system put in place over the last five years, says Kevin Glynn, the company’s chief information officer. “Our salesmen can see our inventory anywhere and data that’s pretty close to real time,” he says. “That’s the first step.” Then comes the harder part: “The gathering of the data is relatively easy. The bigger investment is the analysis of it.”

While Glynn has people to collect the data, he acknowledges that “we don’t have any data scientists on staff. That’s a common concern, particularly among smaller enterprises that struggle to make the case for investing in potentially expensive analytics, when they don’t fully understand how these data systems can improve profitability and customer service.”

Oliver Halter, the Big Data leader for consulting firm PricewaterhouseCoopers, says “small metals producers and distributors are certainly at the beginning of the Big Data revolution. Like most small businesses, their computer systems are not as robust, and traditionally little investment has been made into data warehousing or reporting, let alone advanced analytics.” Halter points out, though, that pursuing analytics has gotten easier—and the necessity has grown clearer. “The investments are dramatically lower than they were 10 years ago,” he says. “The most resistance I see is people not believing that analytical use of data can give them insights,” particularly among an older generation of managers who have long relied on gut instinct and hard-earned experience. They ask, “How can data be smarter than I am? … Asking them to change their behavior because of data is very hard.”

As Halter sees it, the next decade will witness a shakeout between those who choose to embrace data analytics and those who don’t. “The winners are going to be those who realize data is their friend and not their enemy—and are making investments to harness it,” he says. “We will see a percentage of companies that fail because they didn’t recognize the importance of data. It’s going to be a larger number than you might think.”

 

How Data Spurred a Turnaround

Two of the companies leading the charge are major manufacturers, Ford Motor Co. and GE. Each in its own way illustrates how data analytics are reflected in the bottom line.

“We can go through data sets in minutes that would have taken days,” says Michael Cavaretta, a Ford data scientist and technical leader for predictive analytics and data mining. He points to Ford’s Smart Inventory Management System as an example of his team’s ability to mash diverse data points up and down the supply chain—from parts suppliers to buyer preferences—to get the right vehicle into the right dealership on the right day. This includes data on what’s been built and available, what inventory was available for previous sales and what buyers are seeking. If it previously took several weeks for assembly plants to get the information they needed to jigger their production schedules, now they can make changes within minutes.

The result: Five years after a record $14.6 billion loss, Ford tallied its 18th straight profitable quarter in January 2014. 

Operating out of Ford Research and Innovation, Cavaretta is one of some 200 Big Data and analytics experts working throughout the company’s divisions. While he and his team are charged with solving problems, they also generate their own ideas. That’s why, besides needing statistical and computer science skills, he also needs to be a skillful explainer. “If you don’t do a good job at telling the story, then you don’t get to make the change you need,” he says.

Cavaretta sees enormous opportunity to make continuing improvements as Ford expands the use of increasingly inexpensive sensors everywhere from manufacturing equipment to the vehicles themselves. This expanding pool of data will enrich its ability to forecast everything from equipment life cycles to the information that will better serve salespeople and drivers. He admits that the amount of data involved can seem overwhelming, but for manufacturers, the ability to employ and manage this asset “will distinguish who are the competitors and who are the also-rans.”

The lessons from the Ford experience? In addition to efficiently collecting data, companies must have the talent to slice and dice it effectively, be it in-house or a third-party vendor. Just as important, an enterprise’s top management must embrace an often-radical change in how business decisions are made.

 

Connecting the Pieces, Smartly

GE calls its vision for the future the “Brilliant Factory,” which a report from its Global Research team describes as a “21st-century, self-improving manufacturing ecosystem.” This is not simply imagining: In Schenectady, New York, a state-of-the-art plant manufactures newly invented batteries in a massive complex built from scratch. Opened in 2012, it is expected to generate revenue of more than $1 billion by the end of the decade. The Durathon batteries, containing more than 30 patents and used in telecom cell towers and other industrial applications, are half the size of conventional lead-acid batteries but last 10 times longer.

The Schenectady showpiece, about the size of four football fields, offers an impressive collection of machines embedded with more than 10,000 sensors to measure temperature, humidity, air pressure and other machine-operating data. The enormous digital output allows plant engineers to monitor the production process in minute detail, making adjustments in response to information that can be collected up to four times a second.

Stephan Biller, the chief scientist for manufacturing at GE, based at GE’s Global Research Center, admits that the facility generates more data than it needs. “When you build this kind of factory, you want to overdo it,” he says, noting that the plant staff now relies on only a small percentage of the approximately 10,000 variables to make changes. “The reason we put that effort in is that we are still developing the product and improving the quality.”

Biller sees the “brilliant factory” as a window into a new era in manufacturing—a third industrial revolution. Not only will machines “talking” to machines increasingly be able to improve themselves, Biller says, but the new ecosystem will also create more collaborative enterprises throughout the supply chain—to “improve our speed of innovation.” This means engaging engineers and designers through crowdsourcing platforms in product design, and working with supply-chain partners to virtually simulate the manufacturing process with 3-D models before it’s physically implemented in the factory.

These are the kinds of questions Biller is asking: “How can we use the data from all the machines to improve the whole factory? How are we integrating our supply chain with real-time data? How do we take advantage of all our real-time data to operate at a higher level and with greater uptime?”

 

Managing the Shift

On-site data centers and armies of data scientists are almost certainly out of the question for many manufacturers, especially smaller operations and those with older equipment and leadership that questions Big Data’s value. But a growing number of software suppliers provide more limited and less costly solutions. Rockwell Automation and SAS are two that have partnered with companies from diverse industries to provide both pre-packaged and customized services.

North Carolina-based SAS, one of the world’s largest private software companies and a leading provider of advanced analytics, works with companies to support what it calls “business intelligence.” Consider a few examples:

  • Computer maker HP stores data from some 2.5 billion interactions annually, including customer calls, web visits, emails and input from retail partners, which generate service tickets, questions, complaints and suggestions. To better understand its customers and make better data-driven decisions about its investments across the company, HP works with SAS to analyze structured data (such as survey results) and unstructured data (such as online product reviews or other comments that might be found on social media like Facebook or Twitter). Results include a 50% jump in orders shipped and a 42% boost in marketing ROI over three years.
  • United Parcel Service (UPS) tracks data on more than 16 million packages a day, with nearly 40 million tracking requests from customers daily. But its big push into Big Data has come from putting sensors in more than 46,000 vehicles to monitor daily activity (including speed, braking, direction and drive train performance) and redesign drivers’ routes. The initiative, called ORION (On-Road Integrated Optimization and Navigation), relies on map data to reconfigure in real time the drivers’ drop-offs and pickups. Consequently, in the last several years, UPS has cut about 85 million miles off truck routes and saved more than 8 million gallons of fuel.

 

Potential for All

It’s not just the big guys who can take advantage of this data revolution. On a smaller scale, remote equipment monitors and the software to analyze their data are getting at once more sophisticated and less expensive. They’re also becoming more essential.

“Top performing organizations use analytics five times more than lower performers,” according to “The New Intelligent Enterprise,” a 2011 global survey of more than 3,000 business executives, managers and analysts by the IBM Institute for Business Value and the MIT Sloan Management Review. The biggest obstacles are “managerial and cultural,” including not understanding how analytics can improve the business, lack of management focus because of competing priorities and lack of internal talent to manage data systems.

Nevertheless, as Forward’s following story on smaller-scale automation and robotics illustrates clearly, players of all sizes who would stay competitive in a Big Data world require new thinking, new investment and new management attitudes. But then, isn’t that what any successful business must do anyway?


Steven Beschloss is an award-winning editor, journalist and filmmaker. His work has been published in The New York Times, New Republic, Chicago Tribune, Village Voice, Wall Street Journal and Parade Magazine.

To search, type what you're looking for and results will appear automatically