How to Justify IoT Investments
Gary Mintchell | November 15, 2017
Lately, everywhere I travel, the discussion gets around to the Internet of Things (IoT)—sometimes called Industrial Internet of Things (IIoT).
Inevitably, when the term comes up relative to manufacturing and production, it’s linked to “predictive analytics.” That term is just a broader application of the more specific term predictive maintenance (PdM), something this magazine has discussed for many years.
The Emerson Global Users Exchange held its annual meeting in Minneapolis in October. Mike Boudreaux, director of Emerson’s Connected Services business, led a panel discussion featuring Tom Madilao of Chevron Oronite (Singapore); Peter Zornio, Emerson Automation Solutions CTO; and representatives of two supplier partners—Jose Valls of Microsoft and Dave Van Daralaer of AT&T.
Madilao discussed using IoT technology to locate people and steam-trap losses. Tags on plant personnel help managers and first-responders find workers in hazardous areas, in case of an incident. Wireless sensors that monitor steam traps alert maintenance technicians to energy-wasting leaks.
Answering a question from someone in the overflow audience regarding the difference between “Internet of Things” and “Industrial Internet of Things,” Zornio compared it to what GE is attempting to do with Predix—build an industrial-business model where the supplier can sell services based on usage.
A significant part of an IIoT solution is the use of cloud technologies. Boudreaux commented that the issue of using the cloud is over. It’s been proven in critical IT applications.
The next question for the panel was how to financially justify IIoT projects.
Explaining his approach, Chevron’s Madilao suggested picking a small project: Do the engineering and lay a baseline, then do a proof of concept (PoC).
Zornio added, “Target a starter-application area where you have a predefined benefit. For example, something that will clearly save energy. Take a rifle-shot approach rather than a shotgun one.” He further advised avoiding people in analytics and data science who say “just gather lots and lots of data, and we’ll give you answers to questions that you didn’t even think of.” As he concluded, “Most people want answers to questions that they have thought of.”
Boudreaux followed up with, “Do what you can easily measure.”
In other words, make the PoC project easy to explain and measure. When you can show the return, you should be able to get funding for the next level.
If you are attempting an IIoT project to improve predictive maintenance, you’ll likely find one or more barriers to adoption. Emerson points to three major ones: Culture of the company; lack of a clear strategy, including a clear business case; and knowledge of which technologies to use.
By the way, if you have kids, you may need to advise them on which programming languages they should learn. Consensus was: JSON, node.js, Python, R, and bots. EP
Gary Mintchell is an industrial-technology subject-matter expert. He can be reached at email@example.com.