Machine Learning: The Next Failed ERP?

Ken Koenemann | January 2, 2018 | 

If you pay attention to industry or technology news, then you’re seeing a lot of predictions for how new technologies, such as machine learning, artificial intelligence (AI), and solutions for the industrial Internet of Things (IIoT), will transform operations management and the way business is done.

Of course, if you’ve been around awhile, then you’ve heard plenty of similar pronouncements about different technologies over the years. ERP is a great example. It has gone from being a “next big thing” to having a reputation of frequent failures. In fact, Gartner recently argued that the majority of implementations fall far short of their expected value due to a variety of factors, including complexity and lack of a coherent strategy.

Today, ERP solutions offer an important cautionary tale as you consider the current “hot” technology options. With that in mind this post explores some of the keys to success with any machine learning or AI solution implementation (or with any IT solution for that matter), including:

  • What it really takes to see real value from a new solution
  • What investments will drive the most impact (replacing versus updating)
  • How to get started moving in the direction of IIoT and machine learning.

Cool concept versus business realities: seeing real value

Clearly, if you want your business to thrive, you can’t simply ignore technology trends or you could quickly end up losing ground to competitors. Yet early adoption also comes with potential high costs, poor returns and other pitfalls. That, no doubt, is why one of the biggest questions we hear CEOs struggling to answer with respect to technologies like machine learning and AI, is: “How is this going to help me grow revenue or be more profitable?”

The challenge with these technologies is that a lot of the current buzz is conceptual rather than concrete.

Sure some major manufacturers are using sensors and machine learning to improve their operations or products. And most experts would agree that combinations of these technologies have tremendous potential. For example, it’s easy to imagine scenarios where machine learning could help:

  • Understand trends in industrial environments and operations management so you can improve efficiency and profitability
  • Identify trends and patterns in the supply chain that could help reduce materials procurement costs
  • Better understand trends in the retail environments for more tailored product offerings, which could help optimize profitability (although not necessarily revenue).

Identifying trends, however, is one thing. Having the right processes in place throughout the rest of your organization so you can act on them is another. And it’s at this intersection of processes, technology and data that things can get messy and confusing fast.

That’s why, in addition to asking how you could use a new solution to improve your business, it’s even more important to consider your current management approach and ask the following questions:

  • How often do you evaluate process to determine where, or if, there are breaks?
  • Do you have the data, or know how to find the data, to help you identify and understand process issues?
  • Do you have a solution in place that allows you to easily view current real-time performance data and compare it to historic data?
  • If you answered yes to the previous question, then how does your solution help you to improve top-line or bottom-line performance?

The last question is especially important if you are using a dashboard solution, since the vast majority of dashboards are focused on providing “actionable insights.” The problem is that to turn insights into actual business performance improvements, you need to understand more than just whether or not you are hitting targets. You also need context to understand why you are missing targets along with systematic processes that empower your teams to quickly improve the situation using trackable countermeasures.

Getting maximum returns from investments

Once you have the systems, data and process rigor in place to make continuous performance improvements, then it makes sense to start thinking about your next big technology move. At that point, it’s also important to weigh your decision against capital equipment upgrade or replacement needs. Keep in mind that the budgeting and transition planning can be tricky for something like an IIoT and machine learning initiative. For example, although the price of sensor technology has fallen in recent years, if you’re investing roughly 10% in capital improvement projects, then you’re still looking at a 5+ year transition to get up-to-date equipment. With an IIoT initiative, a gradual transition may be ideal because you can start with a small project and take time to learn from the experience and carefully architect the companywide solution. The important thing is to make sure you are able to get the expected value from your early project(s), and that you carefully consider long-term goals before choosing capabilities and platforms or you could end up with technologies that are incompatible with business needs and more ERP-like frustrations.

Building a foundation for success

Given all of the competing considerations, figuring out where to start with an IIoT and machine learning initiative can seem daunting. But it doesn’t have to be. The key is having the right foundation in place to build from by:

  • Focusing on strategy and process: Consider implementing a proven comprehensive management system approach, such as the TBM Management System, which helps every level of your organization stay focused on hitting key objectives.
  • Integrating your data: Find a platform, such as our Dploy Solutions suite of operational excellence tools, that will enable you to more effectively harness all of your existing data and consistently hit your performance goals. Dploy can provide a comprehensive view of your business and it provides a closed-loop system that enables you to drive business performance improvement.
  • Talking to experts: Most solution providers only talk about the end game, but we can help you evaluate the best use of your limited investment dollars. We rely on our 500+ years of operations experience and ability to evaluate projects based on real versus perceived benefits.

If you are considering a machine learning or artificial intelligence initiative but need help getting started, contact us.

Ken Koenemann, VP Technology & Supply Chain
Ken Koenemann

VICE PRESIDENT, SUPPLY CHAIN & TECHNOLOGY
Ken is a 25+ year veteran of manufacturing, operational excellence and supply chain optimization. He believes that it is critical for organizations to address the challenges of strategic planning and use of big data in manufacturing. At TBM, Ken is actively leading the effort to our suite of services to include emerging technologies that improve productivity and convert complex data into information for improved decision making.