Why Industrial IOT is usually a disappointment and how to fix it

Anna-Katrina Shedletsky

If you live and breathe the manufacturing industry, you have spent the last couple of years bombarded by use cases and sales pitches for Industrial IOT products claiming to support your Industry 4.0 revolution, perhaps with AI and other whizbang technologies. With all of the marketing materials being produced with these buzzwords, it’s easy to think that Industry 4.0 has arrived and is within reach.

The data disagrees. In mid-2017, Cisco produced a report of survey results indicating that companies considered 76% of their IOT initiatives failures, and a majority said that IOT initiatives looked good on paper, but turned out to be more complex than expected. In early July, Gartner published its first ever 2018 Magic Quadrant for Industrial IOT, which included companies that provide IOT platforms that work in multiple verticals, such as transportation, manufacturing, utilities, and natural resources. The punchline is that no company crossed Gartner’s bar for execution – indicating an opinion that the products listed weren’t living up to expectations.

What is Industry 4.0? The concept originated in Germany, and describes how initially steam power, and later electric power and mass production assembly techniques enabled the first industrial revolution (from Industry 1.0 to Industry 2.0). Industry 3.0 will incorporate robotics and computers, paving the way for Industry 4.0, which will network everything into a self-learning, self-correcting, system.

As the CEO of a company in this space, I spend a lot of time working with operations and manufacturing leaders across many types of organizations in the manufacturing vertical. These teams exist in various stages of disillusion about Industry 4.0 initiatives, spanning from still seeking the magic technology that will make their biggest headaches go away to shunning anything with a dashboard.

The gap between reality and the marketing is in sequencing: first we must build a solid data foundation, second we must identify and understand the key business metrics we want to drive, and only then can we actually start the transformation of the process through advanced technologies.

Building a strong foundation of data is more challenging than it sounds. The concept of Industry 4.0 is based on the opportunity of networking machines together and aggregating data in the form of dashboards that can inform actions. While networking and data aggregation are part of the foundation, the data itself is often not deeply considered. The easiest stuff to collect and keep is likely not enough to drive major process improvements. Pass-fail data is useful for indicating that there is a problem, but gives no indication of where an engineer should look next; whereas parametric data can not only indicate a problem, but also provide some context that can accelerate the investigation towards a solution. The data that makes the foundation for Industry 4.0 systems must be detailed enough to be actionable, and relevant to the processes being controlled. Relevancy requires detailed knowledge of the process, which makes true multi- or general-purpose solutions like those on Gartner’s list a challenge. Even within manufacturing, this can be a real challenge: controlling a glue process on a consumer electronics assembly line requires different datasets than a stamping process on an automobile body panel line. The final part of the data foundation is that it must be made readily accessible to those who can contribute to problem solving. Typically this means that engineers need to be able to see the data from wherever they happen to be (perhaps in a dashboard form), and also to dig into the details to understand next steps.

The second point is that we need to know what metrics we’re trying to improve, and we need to be measuring them. Whenever the marketing for Industry 4.0 starts to get in the way, I always want to get back to basics: what really matters is better business outcomes. If you can create a better business outcome, then your pilot project for the new technology will be successful. For manufacturing, those are really clear: yield, throughput, uptime, efficiency, and time to market. Budging any one of these metrics even a little bit can create massive returns on investment. The bad news is that there is no magic “raise all metrics” technology that can be slapped on top of your existing processes (sorry). Even with technology, improving these metrics requires the hard work of finding and solving problems.

Building a strong data foundation and having a way to measure and operate against core metrics are difficult, and most companies in the business of designing and manufacturing products don’t have robust systems in place to do so. One of the reasons for that is that change management is difficult: if paper tracking your input, output, and yield has worked to date, it’s difficult to find direct return-on-investment for digitizing it. Another is that piping data around and cleaning it up so that it can be processed automatically is not particularly sexy, and in and of itself doesn’t have direct return-on-investment (though getting it in front of your engineers can enable direct business benefits as they identify and fix issues using it). These investments are table stakes: they must be made first so that it’s possible to layer on the transformation tools that Industry 4.0 is all about. Machine learning models, computer vision techniques, big data analytics, and statistics can create major business outcomes when deployed on an aggregated database of clean data from relevant processes with high enough detail.

In spite of the marketing, most organizations are somewhere between Industry 2.0 and Industry 3.0 – they have mass production capabilities and basic automation, but no aggregated data systems or networking. As they look to understand what Industry 4.0 technologies can do for their businesses, they must keep in mind that in order to be successful, they will need to spend effort on developing a strong foundation of data and identifying the key business metrics that they want to improve. Without this foundation, the rest of the technologies will remain one-off, isolated, and won’t live up to the promises we’ve all been sold.

version of this article was originally published on Forbes.com.

Instrumental is purpose-built for manufacturing to build an aggregated data record and automatically generate insights so your teams can start benefitting today, not in the future. Request a demo here to learn more.