“We don’t think that small manufacturers have realized how easy it is for them to get into this,” he noted. “These questions come up consistently over and over again: ‘How do I add service to what has traditionally been a product-centric marketplace? How do I drive high-margin, high-value services by using the Internet of Things?’”
Happily, in some cases, the answer increasingly involves leveraging data that is already being collected. “Often, customers have invested in those assets already, but they haven’t fully utilized them,” asserted Frank Kulaszewicz, senior vice president, architecture and software at Rockwell Automation, who cites the high-ticket office building elevators produced by ThyssenKrupp as an example.
“They’ve always had factory automation controls that enable technicians to diagnose service issues,” he noted. Now the company is using those sensors to boost the safety and availability of elevators, enabling it to realize a new revenue stream (See sidebar, p. 52). “This is a good example of building a business model around safety and availability, so that beyond the initial purchase price you have an ongoing annuity of providing services afterward,” explained Kulaszewicz. This “servitization of products” has become a growth opportunity for the company, which now provides maintenance for both its own elevators and those of competitors. “They have about 500,000 other company’s elevators on service contracts—those competitors are pretty sad about that,” noted Masson.
In other cases, companies use connected devices to change user experiences and communicate with customers more effectively. Network connectivity, for example, enables Coca-Cola’s smart vending machines to conduct real-time test marketing, track trends and drinking preference and adjust a machine’s selections accordingly. The resulting data helps Coca-Cola understand
customer-buying patterns, deliver more relevant advertising and change the customer’s experience.
For Hydro Electronic Devices, the ability to monitor and analyze 15 variables simultaneously is what customers want. “It’s the algorithm,” said CEO Paul Ludwig, whose company manufactures and markets electronic controls for mobile equipment applications, including software that is used to implement vehicle-control strategies. “Show me a dashboard that tells me where my vehicle is and that tells me what I need to know before I need to know it—the two or three variables that tell when this hydraulic cylinder is starting to fall off. The end decision is where you see the value.”
Being able to pre-empt the downtime machinery failure can involve is a huge plus for his customers. What’s more, over time, patterns and correlations in that kind of data give companies a better understanding of why a piece of equipment fails and what design modifications could be done to prevent that failure.
All of these examples involve “machine learning,” or taking the data insights of the IoT to the next level by mining historical data with computer systems to make predictions about future activity. That activity can be anything from trends and behaviors to performance patterns. “It begins with cause and effect—or predictive maintenance, which is normal math,” said John Dyck, global director, software at Rockwell Automation. “Then you move into machine learning, which is these models that you teach about the machine; and then it takes off from there and gets
smarter by itself and begins making non-causal predictions over time.”