Manufacturing Moneyball: Using Big Data and Business Intelligence To Spur Operational Excellence

The national unemployment rate is 9.1 percent, up nearly four points from the first quarter of 2008. Recent news has not been optimistic. In August, for the first time since 1945, the federal government reported a national net job change of zero. Meanwhile, the global economy sputters.

In these economic times, it is important to examine existing business practices and ask difficult questions: “How can we work smarter? How can we improve our processes? What are we leaving on the table? How can we regain our competitive advantage through operational excellence?” This piece focuses on the manufacturing industry, which is experiencing an industrial renaissance as a result of its application of business intelligence (“BI” or analytics) to Big Data to uncover and leverage data that has remained either (i) hidden or (ii) unappreciated. In this respect, manufacturers that are ahead of the curve are playing what amounts to Moneyball, a strategy first implemented successfully by then-Oakland Athletics’ manager Billy Beane and made famous first by author Michael Lewis and, more recently, by the movie of that name starring Brad Pitt.

This piece refers to a development I coin throughout as Manufacturing Moneyball. Implementing the tenets of Moneyball into manufacturing and other industries is a business imperative. The amount of hidden information and rigorous analysis of it that can create actionable intelligence is overwhelming. Corporations that do not tap into and leverage that resource will be stranded.

Manufacturing Moneyball has ushered in a new era of data transparency and visualization; competitive advantage both at home and internationally; and new business practices. Like the use of other technologies such as cloud computing, which, as we will see, is an integral part of Manufacturing Moneyball, the integration of BI and Big Data is no longer a domain reserved for IT. Those who play Moneyball will enjoy a significant advantage in the manufacturing industry. According to the McKinsey Global Institute:

The use of Big Data is becoming a key way for leading companies to outperform their peers. Across sectors, we expect to see value accruing to leading users of Big Data at the expense of laggards, a trend for which the emerging evidence is growing stronger. Forward-thinking leaders can begin to aggressively build their organizations’ Big Data capabilities. This effort will take time, but the impact of developing a superior capacity to take advantage of Big Data will confer enhanced competitive advantage over the long term and is therefore well worth the investment to create this capability. But the converse is also true. In a Big Data world, a competitor that fails to sufficiently develop its capabilities will be left behind.

Big Data: The Next Frontier of Innovation, Competition, and Productivity (McKinsey Global Institute May 2011).

Manufacturing Today

Before we turn to Moneyball, we must look at today’s manufacturing sector. What one discovers might be surprising given misconceptions about a sector of the economy that is rarely mentioned in the same breath as high-tech or energy, for example. The reality is that manufacturing for decades has played a central role in our economy and still does.

The United States has the world’s largest manufacturing sector, with a 20% market share since the 1980s. One in six American jobs is still in or directly tied to manufacturing. The latter group is the result of manufacturing’s tremendous multiplier effect, as described below. According to the National Association of Manufacturers, because of increasingly sophisticated technologies and the processes it employs, U.S. manufacturing relies on a more educated workforce and pays higher wages and better benefits than other sectors. U.S. manufacturing alone would comprise the eighth largest economy in the world. In current dollars, manufacturing GDP makes up over 12% of total GDP. More important, its real impact on the U.S. economy can only be measured when we consider its multiplier effect when it comes to jobs and competitiveness in other sectors such as high-tech, telecommunications, wholesale and retail, finance, and accounting. According to the Milken Institute, every job in manufacturing generates more than 2.5 jobs elsewhere in the economy. See Ross DeVol & Perry Wong, Jobs For America: Investments and Policies for Economic Growth and Competitiveness (Milken Institute Jan. 2010). Each of the aforementioned sectors depends on a strong manufacturing base.

The figures cited above should not leave us with the impression that everything is rosy in the manufacturing industry. It is not. Numerous factors challenge its health. Rising external costs in our global economy may well top the list. Domestic factors include corporate tax rates, rising health care costs, regulatory compliance, and energy prices. Those costs add a stunning 18% to manufacturers’ costs relative to our major trading partners.

This leads us back to the questions I posed above:  “How can we work smarter? How can we improve our processes? What are we leaving on the table? How can we regain our competitive advantage through operational excellence?”

Manufacturing, Big Data, and the Need For Business Intelligence Infrastructure

The manufacturing industry for years has moved to answer these questions. Among its most important answers has been the application of business intelligence to Big Data—large pools of data that can be captured, communicated, aggregated, stored, and analyzed—to mine for hidden patterns and insights that lead to efficiencies. This is the essence of Manufacturing Moneyball.

As I have written elsewhere about Big Data, the latest IDC Digital Universal Study reports that more than 1.8 zettabytes—10 to the power of 21 bytes—of information will be created and stored in 2011 alone. Experts predict that by 2020, there will be a 42(x) growth in data from 2009. These figures are as astonishing as they might seem unmanageable. Yet leaders in the data storage market are in fact managing it, especially with the use of cloud computing, which scales to allow BI to leverage unprecedented commercial computing power to analyze high-quality, high-frequency Big Data. Manufacturing Moneyball taps into precisely such data. Doing so is not a luxury. It is now clear that “[l]arge-scale data gathering and analytics are quickly becoming a new frontier of competitive differentiation.” See Jacques Bughin, John Livingston & Sam Marwaha, Seizing The Potential of Big Data (McKinsey Quarterly Oct. 2011).

In assessing the value of investing in BI-, data-focused infrastructure cannot be overstated. Even the largest manufacturers will only realize and maximize this opportunity to create business value once they recognize the importance of—and invest in—the convergence of analytics and Big Data, as well as the cloud infrastructure upon which it depends. See The U.S. Sustainable Business Market: Spending Patterns in (IT) Solutions and Architectures Currently Supporting Sustainability and GHG Management by Manufacturers (Gartner Group Report. No. G00213092 Aug. 30, 2011).

It is not easy to identify the “correct” data to analyze. Even technologically savvy organizations have critical data that resides in antiquated legacy systems that may not be integrated (e.g., by enterprise application integration). Much information is siloed and resides in isolated datasets with different custodians or in enterprise resource planning (ERP) tools, says Sean Robinson, Global Industry Manager of GE Intelligent Platforms and an expert on using BI in the manufacturing sector. Yet without significant investment, manufacturers will miss business-critical opportunities to achieve operational excellence by marshaling and leveraging one of their most important corporate assets: information (data). The next section examines these opportunities in detail.

Manufacturing Moneyball

The heart of Billy Beane’s revolutionary baseball strategy was finding hidden information about players’ performance that other teams overlooked or undervalued. Beane and his young statistical mastermind, Paul DePodesta, realized that a player’s batting average alone was a poor measure of his value to a team. Rather, all teams must score runs to win, and in order to score a player must first get on base. This shift in focus to on-base average changed baseball.

The manufacturing equivalent of Beane’s strategy is (i) to find hidden data—and specifically real time data at both aggregate and granular levels; (ii) to apply advanced computational methods to mine that data; (iii) to understand the multivariable data’s interconnectedness; and then (iv) to optimize traditional manufacturing processes accordingly. Linda Onnen, Global Marketing Director of GE Intelligent Platforms, told me recently by phone that Big Data allows for precisely the sort of multivariable analysis required to identify those links and then analyze them to generate actionable intelligence.

A manufacturer’s ability to achieve these goals depends on the same two variables that Beane and DePodesta used to evaluate talent and turn around a franchise: good information and analytics that lead to actionable intelligence. Gartner writes:

The quality of the[se] outcomes is strongly linked to the quality of the data being collected and the manner in which it is stored, analyzed, visualized, and converted into meaningful and valuable sustainable business intelligence (BI).

The multivariable analysis to which Onnen refers depends on at least three factors:

1. Data transparency that allows data from different manufacturing functions to be integrated. This allows executives to form a holistic view of the processes never before possible.
2. Process visibility that allows managers to see how processes unfold as they happen, which allows for real-time adjustments.
3. Data visualization. As analytics are applied to Big Data, the output must be analyzed mathematically and represented visually so as to allow end users such as plant managers to actually see the hidden data and its value. This is especially important in light of the fact that the data is dynamic in real time.

Multivariable analysis allows organizations to see both their informational assets and shortcomings. As Onnen states:

Manufacturers can greatly benefit from an integrated "big picture" view of their operations whereby they can measure environmental factors, for example, alongside other operational metrics to gain visibility into where they may be disrupting stable processes, or conversely, where they are accepting inefficiencies simply because they do not realize it.

Manufacturers with better information realize continuous improvement that allows them to focus their efforts on margin recovery, rather than on capacity improvement alone. These improvements take many forms, including:

1. Margin Recovery. Savings are regularly found in reduced materiel costs and system capacity recovery—for example, yields in packaging.
2. Product Quality and Safety. Teams are better equipped to understand and eliminate the true root causes of product risk. They can thereby actually address and fix these shortcomings at the start of processes rather than merely discarding low-quality output based on post-production testing. Product quarantines can be established based on specific, objective data rather than subjective approximations.
3. Eliminating overlapping investment and personnel support. Identifying overlap should result in the proper allocation of resources—both people and technology. Yet it is critical to emphasize that the human element of the manufacturing processes cannot be decoupled from even the most advanced analytics. On the contrary, is essential to BI, as discussed below.
4. Measurable ROI. Moving from the sort of traditionally unstructured approach that characterized previous manufacturing reporting processes to a modern solution built for data collection and analysis enables management to develop consistent, new methods that can result in greater efficiencies, yields, and production flexibility. Such gains allow manufacturers to take new products to market up to 30% faster than before, a remarkable achievement with significant effects on profits and losses.
5. Collaboration. The synergy of BI and Big Data provides both macro and granular views of information that allow management, operators, and engineers to work together based on quick feedback in a data-driven environment. This is not the siloed manufacturing industry of the past.
6. Monetizing Assets. Corporate attitudes with respect to monetizing corporate assets such as intellectual property have changed dramatically over the past five years. In manufacturing, this move towards using BI to mine Big Data has shifted the view of those assets and the systems that generate them as profit-enabling centers rather than just insurance and a cost of doing business. This is a profound change with significant effects on an enterprise’s bottom line.

The Relationship Between Human and Business Intelligence

The relationship between technology-assisted processes and the people needed to make them succeed has spurred vigorous debate in numerous businesses, e.g., the electronic discovery process in litigation. Let’s cut to the chase. No matter how advanced technology may be—in this case highly advanced mathematical algorithms that produce business intelligence—it is a costly mistake to divorce the processes that it drives from the specialized, tacit human knowledge and expertise that exists in any enterprise.

Decoupling the two is a serious step in the wrong direction. This is not to say that relying on humans alone to guide manufacturing processes suffices. It does not, and that shortcoming is the very focus of this article. Likewise, analytics alone cannot be the sole catalyst of this industrial knowledge. Just as corporations have discovered hidden, siloed data, so too must they tap into and then incorporate into their BI the extraordinary amount of tacit knowledge that resides (often hidden) in their workforce. Sean Robinson describes it this way:

To some degree you look as much as possible to leverage data that is coming out of equipment or sensors—data that is inherent to the process. This is the basis for BI that should guide process improvements and standards. But the way that people interact with that process tells you a lot about what they know and believe to be true. So when you have a skilled, experienced operator who overrides temperature settings in a cooking process or who shortens a cycle time, you have to ask “why”. The key is to be able to sit down with those operators to understand why they’re making the choices that they’re making, to capture it, and then to combine their human expertise with the underlying objective data in order to create new operating procedures and processes.

Capturing this knowledge is even more important when one considers natural employee attrition and retirement. Expertise that walks out the door cannot easily (if ever) be recovered and can significantly affect operations and business in the long term.

Expertise gained through BI must be “socialized” among managers, engineers, and other workers so that it may be transferred to business (manufacturing) processes. In this respect, one must consider the types of workers one finds in manufacturing today to dispel the common visual that people have of highly repetitive, task-oriented types of jobs and replace it with the reality of manufacturing’s high level, analytical problem solvers. I discuss this above. Robinson further explains:

Operators today are not the stereotypical industrial types anymore, and working in a modern factory is intellectually engaging, not just rote work. It’s not a job you do because you have no other choice in life. Modern manufacturing attracts workers and analysts who might otherwise get a graduate degree in math but find it just as rewarding to be part of a team that’s doing process analysis and continuous improvements, whether that relates to redesigning transmission lines for a major car manufacturer or making baby formula in a food plant.

Manufacturing Moneyball’s Return on Investment

Manufacturing Moneyball has a measurable return on investment (“ROI”). Moving from traditionally unstructured approaches often found in older processes to a modern structured solution that is designed for data collection and analysis to enable enterprises to develop new, consistent processes can result in 14% higher production yields.  Software empowers teams of operators, engineers, and management to work collaboratively and identify areas of improvement in a data-driven culture predicated on quick and accurate feedback. With higher yields, products can go to market more quickly. Enterprises can tighten quality specifications. Margins are recovered. Savings are found. Materiel costs decline.

Each of these returns on investment on Manufacturing Moneyball give companies competitive advantages that were not possible before. Robinson recounted his experience with the Vice President of Supply Chain of a Fortune 100 company who had implemented data-driven analytics in 80 factories for two years. In a meeting with 120 IT managers and employees, he translated Moneyball into concrete terms: their initiatives have resulted in an $80 million reduction in hard costs and a 4-point net margin improvement for affected business units.

Those are extraordinary figures when one considers how those savings can be used:

1. reinvestment in the company’s workers and product lines;
2. job-promoting expansion;
3. as a buffer against price pressures from competition and other market factors; and
4. proper capital allocation based on new business plans and process strategies.

Perhaps most interesting, newly found efficiencies can help companies on the supply side both accelerate their speed to market and provide more accurate data to their buyers, which results in improved customer loyalty and credibility within the supply chain.

Conclusion

The manufacturing sector has come far in its use of data and analytics, but not far enough. MIT Professor Eric Byrnjolfsson wrote only this month: “Too many managers are not opening their eyes to this opportunity and understanding what Big Data can do to change the way they compete.” Eric Byrnjolfsson, Jeff Hammerbacher & Brad Stevens, Competing Through Data: Three Experts Offer Their Game Plans (McKinsey Global Institute Oct. 2011). Manufactuers who do not have a long-term holistic vision will be at a significant disadvantage vis-a-vis their competition. They will find it necessary to keep many jobs abroad rather than bring them back to local markets and more highly skilled workforces. Critical information will remain hidden, making it difficult if not impossible to drive improvements or increase production without adding equipment or people. They will become seen as less reliable supply chain partners. The parade of very real horribles goes on. At the end of the day, they will become dinosaurs.