Jim Euchner, From the Editor, Vol 61.5
Give ordinary people the right tools, and they will design and build the most extraordinary things.
The Internet of Things (IoT) has taken on many meanings. In manufacturing, it means instrumenting and informating factories so that data can be used to improve quality and productivity. In logistics, it means a unique identifier for individual items so that supply chains can be made more intelligent. In new product development, it means the development of smart, connected products that provide information about their state so that information can be used to improve the operations the products support. It can also mean the use of information to broaden the traditional design space for a product. Finally, for manufacturers, IoT means the proliferation of data-based business models, especially those that sell products as services.
Today’s IoT implementations are mostly part of what Neil Gershenfeld, who helped originate the concept, would call the BITNET of Things. The applications generally involve sensors connected in some way to a central controller, which does the analysis and sends alerts or instructions. The devices themselves are not directly connected to the Internet. To Gershenfeld, being on the Internet directly, as opposed to attached to a server on the network, is a crucial distinction, one that affects the potential for innovation. In his view, IoT “means IP [Internet protocols] embedded in devices” (Gershenfeld and Euchner 2015Gershenfeld, N. G., and Euchner, J. 2015. Atoms and bits: Rethinking manufacturing. Conversations. Research-Technology Management 58(5): 16–21., p. 16). In this model, devices can communicate directly with other devices, with users, and with the wider network, making innovation possible without access to the central server.
As an example, there are commercial applications that use sensors on tires to gather information about tire pressure and then use that data to predict and prevent roadside failures. Now those sensors do little more than send a stream of data to a central server; applications on that server analyze the data and relay the results to others—users, service providers, fleets—as directed by the central algorithms. In Gershenfeld’s vision, this limits innovation. If, instead, those devices had their own embedded IP capabilities, anyone could innovate with and around them:
The tire might want to talk to an electric motor driving the wheel to convey torque information; it might want to talk to the other tires to convey information about relative loading on the car; it might want to talk to the UI for the driver to let the driver know the tire needs attention; it might want to talk to the highway to convey information about where on the highway, for example, load is being applied. Beyond that, it might want to send information to manage the inventory of tires as part of a work process. (Gershenfeld and Euchner 2015Gershenfeld, N. G., and Euchner, J. 2015. Atoms and bits: Rethinking manufacturing. Conversations. Research-Technology Management 58(5): 16–21., p. 17)
With true IoT, as Gershenfeld conceives it, innovation can happen at the edge—at the device itself. Anyone can access the device’s IP stack, so anyone can innovate, which creates unimaginable potential to increase the power of these devices and accelerate the introduction of new services.
This future is still on the horizon, but there is a lot of exciting work going on today. This issue includes articles on different aspects of current IoT practice as it moves into the mainstream. In “Smart Factory Implementation and Process Innovation,” David Sjodin and his coauthors discuss a maturity model for implementing smart factories, which they define as factories that use data from the factory floor “to support dynamic adaptation and maximize efficiency.” Based on their study of five factories in two automotive manufacturers, they identify and discuss the challenges associated with implementing the smart factory concept in practice.
In “A Straightforward Route to Sensor Selection for IoT Systems,” Paul Jones and his collaborators discuss a three-sieve process for selecting sensors for IoT applications based on sensor requirements, environmental constraints, and economic factors. Their approach optimizes the final sensor selection for the particular application, which is a common need at this stage in the evolution of a new technology.
In his Point of View piece, “Substitute or Synthesis?,” Ulrich Lichtenthaler considers the interplay between artificial intelligence and human intelligence in interpreting IoT (and other) data. He argues that there are four categories of AI applications and that the most potent applications will be built on a partnership between human and machine. Both of these discussions will help practitioners implement IoT applications based on the centralized model (though one suspects that AI partnerships with human intelligence may require a more intelligent edge).
Finally, in “Improving Usage Metrics for Pay-Per-Use Pricing with IoT Technology and Machine Learning,” Timon Heinis, Christoph Loy, and Mirko Meboldt discuss advanced services business models. These models generally include a revenue model based on usage of the product (like the Power by the Hour model pioneered by Rolls Royce). The authors of this article argue that usage metrics that better reflect the true costs of using a product can be developed with the help of data from IoT. Their belief is that these more sophisticated metrics have the potential to broaden the scope of businesses for which servitization makes economic sense.
This issue’s Conversations interview is with Steve Blank, godfather of the Lean Startup movement. Both his interview and the profile of Gary Gray, vice president of product management at Indian Motorcycle, emphasize the importance of customer insight in innovation. Technology—even a technology as powerful as the IoT—is just an enabler. Steve Blank, who has likely observed the dynamics of more Lean Startup initiatives than anyone else, discusses the collection of capabilities that has evolved into Lean Startup and provides insight into the successes, failures, and prospects for the approach in large corporations.
The IoT is exploding today. Visionaries see a future in which trillions of connected devices serve billions of connected people. As exciting as today’s applications are, the future of the IoT may be even more compelling. Very few of today’s IoT devices are full-fledged citizens of the Internet. When these devices start to incorporate simple IP stacks, the IoT could give rise to the same sort of distributed innovation that the Internet has, enabling anyone to innovate at the edge. What is happening with the IoT today may be analogous to what happened pre-Internet with networked computing. Companies implemented (and controlled) entire applications, from the edge to computational centers. When the connections were opened up with the IP architecture, innovation exploded. Perhaps the future of IoT will follow the same path and lead to the same kind of generative growth as the Internet itself.