Skip to content

Making Hay With Reams of Data—From Supply Chain Players

By Karen Haywood Queen Contributing Editor, SME Media

Need help with condition monitoring outside the factory? It’s just the start.

Indranil Sircar, CTO for the manufacturing industry at Microsoft, calls “cognitive supply chain” the future. “It will make supply chain agile, responsive and flexible in developing disruptive capabilities, new business models,” he said. He spoke at Hannover Messe in 2016.

Within manufacturing, a shift has begun to extend big data analytics and condition monitoring beyond the factory door to both ends of the supply chain—to customers, other end users and to second-and third-tier suppliers. Some manufacturers also are expanding the amount and type of data analyzed. They are reacting nimbly to new tariffs, weather forecasts and moving fast to mitigate or prevent problems. The companies moving in this direction are seeing better planning, better inventory control, less down time and fewer recalls.

“Manufacturers need a dependable supply chain; they need to know that the materials they require are arriving when they need them and in a condition that they can use them,” Ed Cuoco, VP of analytics at PTC, said.

“Cognitive supply chain is the future,” Indranil Sircar, CTO for the manufacturing industry at Microsoft, said. “It will make supply chain agile, responsive and flexible in developing disruptive capabilities, new business models.”

“Leading companies have realized that it is time to drive that change, rather than be driven away by it,” Bill Boswell, who markets MindSphere for Siemens PLM Software, said. “Digitalization is changing the industry landscape by opening the door to more advanced analytics and data collection directly from the shop floor, be that the manufacturer’s production floor or its suppliers. Digital technology is leveraged today to transform many businesses.”

The way it works is quite simple, Hans Thalbauer, senior VP at SAP, said. “You know where you have your suppliers. Then you also know about their suppliers, where they are building the components of the product. If you have all this information, you now have a complete network.”

Better artificial intelligence, machine learning, modeling, Digital Twins, blockchain technology, improved hardware and software, and better more affordable data storage and real-time telematic data are helping power this change—as are cloud architecture, an open cloud-based IoT operating system, seamless connectivity to devices and options of multiple infrastructure as a service.

Also helping drive the change are edge computing and new devices with transient storage and services where operational asset data is aggregated with data from the Digital Twin, Boswell said.

Analytics is used with these Digital Twins to provide actionable insights to optimize the predictive Digital Twin characteristics and improve operational performance, he said.

With adoption of Industrie 4.0, manufacturers are looking to apply new capabilities and innovation in traditional supply chain model and that’s also driving the shift, Sircar said.

“The first stage in that transformation journey is about building a digital supply chain foundation by leveraging the cloud to connect, automate and visualize an end-to-end view of business—whether planning, inbound supply, production or capabilities in sales and distribution,” he said. “With Big Data, IoT and machine learning, we are seeing manufacturers move from reactive to predictive capabilities.”

Revolutionary changes in pricing also are helping fuel the change, Boswell said: “Subscription-based and demand-driven pay-as-you-go pricing models are not only are making the adoption of digital technology more feasible but also fundamentally changing the business models of these manufacturers on how they support and service their products in the market place.”

The transformation, however, is still in its early stages.

Some companies have achieved a shift but have not adapted all the way through the supply chain. Other companies are collecting information but are not using and analyzing that data effectively.

“So far, most companies have achieved a collaboration with their supplier,” Thalbauer said. “But only a few have visibility with their supplier’s supplier.”

That understanding and visibility past the first tier is important.

“If your supplier’s supplier is in Japan, and you see the weather forecast determine this company is going to be impacted by a typhoon in three weeks and will likely shut down, this has an impact on you,” he said. “Without this visibility, you continue to produce until you are hit with the situation where you can’t get the supply anymore. If you have this visibility, you can decide before the typhoon hits to order a few more components so you can overcome the situation until the plant is up and running again.”

Merely collecting the data isn’t enough. Treating data collection as a goal in and of itself isn’t enough either.
Instead, manufacturers need to identify a problem to solve and focus on how better data collection and analytics will solve that problem.

“Most manufacturers are trying to use Big Data and analytics,” Cuoco said. “However, there is a lot of variability in how successful these efforts are. The successful leveraging of Big Data and analytics starts with properly understanding the business needs, desired outcomes and operational challenges/opportunities. Big Bata and machine learning are treated as tools applied to those challenges. Companies struggle when they lose focus on the business outcome and instead take an approach that treats Big Data and analytics as ends in themselves. They invest in data collection and analytics with no context and assume useful insight will present itself.”

Much of this information being collected—the weather forecasts and the identity of the supplier’s supplier—is not new. But the data is only recently being mined for insights and by only a few companies, Thalbauer said.

“Even today, many companies do not have that type of information,” he said. “They don’t understand the benefits.”

The companies that do understand are reaping rewards—by, for example, moving quickly to change suppliers when new tariffs on Chinese steel and aluminum would impact their supply chain and ultimately their end product.

“You want to understand the impact on the product you are producing,” Thalbauer said. “You want to know if your company has components produced in China that will be impacted by these new tariffs. If you have the information about your raw materials, you can understand the impact and figure out if you need to relocate your supply chain.”

Companies that are really connected to their supply chain are able to react nimbly, he said.

“Immediately after new tariffs were introduced, these companies were fast in adjusting their supply chains,” he added. “It’s quite interesting how fast they were able to adjust and redo their supply chains.”

Manufacturers now have a better understanding of market shifts that are taking place because of regulations and the impact of regulations and policy changes in various countries, Sircar said. That knowledge enables them to do intelligent planning, he said.

Beyond condition-monitoring along the supply chain, manufacturers who are leveraging the most benefit are integrating insights into product design, product lifecycle management and other processes across the entire manufacturing enterprise, Cuoco said.

Companies leveraging these capabilities also have seen significant reduction in working capital by having the right inventory at the right time. At some companies, better data analysis along the entire supply chain has enabled the company to move from weekly orders to 48-hour ordering, Sircar said.

“We have seen other improvements, including a significant reduction in working capital needed by taking inventory at the right time,” he said. “That leads to a reduction in capital reserve needed for replacement parts. On the IT side, we have seen significant reduction in man hours needed for data preparation.”

Manufacturers also are applying Big Data analytics to gain insight on predictive returns, Sircar said. They utilize machine learning on data sets, such product testing and customer experience telemetry, to improve product quality, better forecast returns and improve overall customer interactions.

One Siemens customer, a large computer equipment maker, collected data from disparate systems, including component suppliers, peripheral suppliers, failure analysis, IoT devices, field failures, returns and repairs and customer call logs. By analyzing this data, Boswell said, the company:

• linked commonly dispatched parts and required fewer dispatches;
• resolved problems detected across the supply chain;
• optimized test processes and
• predicted field failure of components and systems.

Based on that data and analysis, the company was able to push solutions to customers before problems occurred, he said.

The customer also realized savings of $15 million annually in warranty and excursion costs to resolve problems, he said. Overall, the company reduced by 91 percent the time it takes to resolve customer issues.

Another Siemens customer, a detergent factory, implemented centralized analytics, visualization of cleaning plants and products, as well as performance management and predictive maintenance, he said. With the resulting data and analytics, the company was able to optimize use of water, energy, detergents and additives. The company reduced downtime by 10 percent and reduced consumables by 6 percent.

“The real ROI comes when companies improve their operations and logistics while also integrating that knowledge into decisions around product design, cost to manufacture, even in the commitments made to customers,” Cuoco said.

For example, sometimes the part that appears to be the least expensive actually costs more when all factors are analyzed.

“When designing a product to be manufactured, there are often different options for where to source parts and sometimes the cost of the part isn’t really the right metric,” Cuoco said. “If a given component is cheaper to buy but is 20 percent more likely to have recurring shipping delays, then the ‘more expensive’ part suddenly becomes more cost effective.”

Manufacturers that are gaining the most benefit from analyzing big data also are treating that data and all aspects of data collection as a critical business asset, Cuoco said.

The data shouldn’t simply be collected.

“The condition of the data and the analytics should be constantly monitored,” he said. “There needs to be a process for deriving insight, accuracy and precision, for refining the logic, for deciding when to add new data and when not add new data.”

Manufacturers also need a review process to take action when the process does not succeed, Cuoco said.

“Data governance, investment in infrastructure and an understanding that the use of analytics is an ongoing effort are all important and no different from the processes that support critical equipment or business functions,” he said.

For manufactures moving to the product-as-a-service arena, supply chain analytics becomes even more important.

“Some manufactures are starting to change their business model from a producer of things to a producer of outcomes,” Cuoco said. “They don’t want to build engines; they want to sell their customers hours of thrust. If you want to do that kind of business transformation, you need to have an extremely refined understanding of your supply chain and logistics, both in terms of operations and risk exposure.”

  • View All Articles
  • Connect With Us

Related Articles

Always Stay Informed

Receive the latest manufacturing news and technical information by subscribing to our monthly and quarterly magazines, weekly and monthly eNewsletters, and podcast channel.