Not only is my golf game terrible, it also hasn’t improved in 20 years. I’ve learned to accept it. Still, improving shouldn’t be so hard. It seems like no amount of new equipment or practice makes a difference. Perhaps what I need is a change of perspective.

A friend advised me to accept the inherent variability that comes with the game. Instead of planning for the perfect shot, understand the variability of every aspect of the game – your swing, the ground conditions, that northeasterly breeze. Then, based on that, don’t aim where you want to hit — aim where you’re most likely to minimize the risk of disaster.

This advice also has relevance for supply chain management, where many leaders aspire to make it an exact science. While this may be laudable, eventually the laws of diminishing returns kick in.

Rather than trying to “perfect” the supply chain, maybe a better approach is to focus on risk – and then build out the capabilities needed to manage it. 

Sorting the supply chain: The dimensions of risk

But first, what are the three key aspects of supply chain risk:

• Risks you faced in the past: Past risk is all about analysis to tell you something about the here and now. How did past risk scenarios play out relative to what you anticipated? Was the risk hidden or ignored? By analyzing past risk, you can apply lessons learned to current circumstances.

• Risks that you may face in the future: Future risk is about prediction. Unfortunately, the sources of future risk are many – a new pandemic, geopolitical tensions, trade wars, inflation, climate disaster, raw material, and labor shortages, changing consumer demand, and competitive moves. This list goes on. What is the potential for any of these to impact operations? And what can you do today to protect against the risk if it materializes?

• Risks that are happening right now: Present risk is about responsiveness. How big is the risk? Should we respond now or sit tight? Insight into past risk will help you make better decisions, as will any plans you may have to deal with future risk. Then, with the right capabilities in place, you can better understand the present risk and make informed moves.

But the question remains, how exactly do you evaluate risk?

Bayes’ Theorem and the importance of relevant information

On YouTube, I came across a video by Julia Galef on Bayes’ Theorem about calculating the probability of an event based on an understanding of prior conditions. Rationality, it turns out, is not so much about knowing facts as it is about knowing which facts are relevant.

Galef illustrates this point with a scenario: imagine an introverted student crossing a college campus with books in hand and eyes averted. Is this student more likely to be a math PhD candidate or a business major?

Most people choose the math student because when it comes to introversion (for this thought experiment), 75% of math PhD candidates fit the bill while only 15% of business majors do.

But consider this: for every math PhD candidate, there are 10 business majors (a 1:10 ratio). If there are 20 math PhD candidates, there are 2,000 business majors. Do the math and you’ll see that 15 of the total pool of math PhD candidates are introverted – versus 300 for business majors. Clearly, the shy student crossing campus is much more likely to be studying for a career in business.

The context-aware supply chain

Bayes’ Theorem shows us that we humans aren’t always the clearest thinkers. As a result, for effective supply chain management we need to get systematic –leveraging whatever tools and artificial intelligence we may need to understand context and make better decisions.

To succeed, follow through on three courses of action:

1.     Connect every process: 

Integrate processes for design, planning, manufacturing, logistics, maintenance, and service. When all silos work together and share information, you can dramatically increase visibility. This can help you better understand what the relevant facts are.

2.     Collaborate with your ecosystem: 

Create dynamic, digital connections across all your suppliers, contract manufacturers, logistics partners, and services providers. The result is full visibility into every partner’s capabilities, capacity, and performance.

3.     Contextualize every decision: 

With connected processes and robust collaboration, you can understand the context for decisions. Now, you can work together as one business to avoid the gaps, inaccuracies, and mistakes that slow progress and increase costs. And with access to AI and other intelligent technologies, you can detect patterns that humans cannot. The result is better decision making.

Family-owned business Box Print has been manufacturing high-quality graphic print bags and packaging supplies for more than 60 years. Using a digital and collaborative process approach, Box Print can better anticipate demand, even before a customer needs products, helping it procure raw materials and start production at an earlier stage than before. In an industry where machine setup is the principal cost, accurate forecasts also enable larger production runs with fewer mistakes, helping Box Print offer more-competitive prices. And cash flow is improved with larger runs, lower delivery costs, and more-efficient stock management.

Go for supply chain risk resilience, not perfection

At a time of growing complexity and uncertainty, supply chain perfection is perhaps unattainable – or at least not worth the effort. Few if any plans go as initially conceived.

A better approach is to focus on mitigating the risks that could derail operations. With the right frame of mind and the right capabilities in place to help you execute, you can dramatically improve your supply chain game.

Learn why supply chain professionals now rate risk resiliency as their top priority – and what they plan to do about it. Check out the recent study.

Interested to talk to someone at SAP? Click here.

First Published in The Future of Commerce.

We will begin building our 4^th industrial era automation system at the foundation; Level 0. Level 0 is the base of the automation hierarchy where we interact with the process and get data from sensors.

Data is only as good as the source. If you do not have enough resources to maintain your field devices, the data they generate will be misleading or useless. Electrical and Instrumentation (E&I) is the department responsible for this maintenance.

E&I is often understaffed and overworked. It is very important to prioritize their work. In a circular manner, data can help maintain data.

Part of the responsibility of E&I is to maintain motors and valves, another part is to maintain the sensors that generate data. If the work for motors and valves can be managed more efficiently, it will allow maintenance of sensors to be more efficient and effective.

It is very likely that your motors and valves are automatically accumulating usage data, such as motor run times, number of start/stops, and number of valve strokes. It has become much easier in PLCs and DCS controllers to implement motor runtimes, number of start/stops, and number of valve open/close cycles. This is important to know because motors and valves wear down with usage and require maintenance in proportion to how much use they get.  

Obviously, you do not want equipment failing when you need it. You would like to plan your maintenance. This is especially critical when many mills have fewer maintenance personnel to cover growing inventory of equipment. You cannot afford to waste labor hours on equipment that has not been used, and you need to ensure that high demand equipment continues to perform.

One approach is to use your calendar. When was the last time maintenance was performed on this unit? The problem with that approach is that the unit may have been idle that whole time, and not require maintenance. Your technicians likely are stretched thin and it is important to have them effectively utilized, so having them service equipment that doesn’t need it is costly in several ways. On the other extreme, it may have been in use more than you expect and is overdue. It can fail before you expect it to.

Assuming you already have usage data in your control system, it is a matter of making that data useful. It is not useful if nobody knows about it.

1. A low cost first step could be to implement alarm limits on usage data. It is not that important to know the usage until it reaches a limit that indicates maintenance is merited. A warning alarm followed by a high priority alarm is a reasonable approach. The challenge with implementing any alarm is to follow good practices such as the ISA 18.2 alarm management standard. Filling up an alarm journal does not help. Operators need to know what actions are required to respond to alarms and keep the alarm journal from being filled. An additional issue is that operators are not the ones that need equipment usage information. It is the maintenance manager that needs this information. If an alarm triggers an email to the maintenance manager, this can help.

2. A better way is through connectivity. Your maintenance manager should already have a computerized maintenance management system (CMMS). This is the system that has the inventory of equipment and gives the maintenance department the information they need for planning what to inspect/fix and when. Some of these systems just give you dates to know how long it has been since the last maintenance on each piece of equipment. A better way is to schedule based on usage. It is impractical to perform data entry of the usage data from the control system to CMMS. Realtime connectivity of usage data into CMMS is the way it is done in a connected facility. Not every CMMS supports this connectivity. Therefore, the cost of implementation is to get a CMMS with connectivity if you don’t have one, then once you have this capability it requires configuration to get the usage data from the tag in the control system to the matching equipment in the CMMS. Once implemented, the maintenance manager knows that the CMMS can be used every day to present realtime planning information.

3. The next step would be predictive maintenance with machine learning. This would use high frequency vibration or realtime process data to identify abnormal equipment conditions. We all know that there are times when equipment fails much earlier than we expect it to. If the CMMS can tell the maintenance department that something does not look right with a motor or valve in realtime, it can prevent a much more costly and dangerous failure.

In this third approach, there are two techniques being used in industry:

• High speed sampling of vibration can yield a failure signature. As the equipment is running and in production use, the system can be continuously sampling and comparing the spectrum to known failure patterns. An advantage of this approach is that when there is a match it can tell you exactly what is wrong and what to fix. The disadvantage is that if there is a problem that vibration does not identify, you won’t know about it.

• There is instrumentation other than vibration that pertain to the asset. Amperage, inlet and outlet pressure, flow, winding temperatures, and other measurements can be combined into a Machine Learning model. This model can identify what normal patterns are, and therefore notify when the asset is behaving abnormally. This is a more comprehensive use of data to catch failures that vibration alone might not be able to predict. A disadvantage is that when there is a prediction of abnormal behavior, you may not be able to tell the technician what to fix. Also, the caveats of Machine Learning that I covered in my prior articles still apply.

Next month we will continue this discussion by looking at sensors.

I inherited terrible teeth. As a teenager and young man, I was always self conscious about my teeth. When I could afford it, I had braces installed. They were on for thirty months.

As life went on, I continued to spend money perfecting my teeth. About thirty years ago, upon a recommendation, I started going to probably the best dentist in the region where I live. He did many good things to satisfy my perfect smile obsession.

We never quite got there. That wasn't his fault, it was a change in attitude on my part. As I got older, I realized that what I really wanted was teeth that, just as they drop me into that final box, they fall out. If they lasted longer than my demise, too much money has been spent on teeth.

I told my dentist this and he looked at me blankly. Finally, I left this dentist, for he was a perfectionist, and I no longer needed perfection in this area of my life. He wouldn't change, so I had to make a change on my own. My new dentist now checks after the hygienist cleans my teeth, declares me "Good to go" and that is it. We are both happy, for we are achieving the quality level I desire, which is no longer perfection.

This is what Martin Barkman is talking about above. I doubt that I am risking much with the current state of my teeth, and I accept that they are functioning and people don't retreat in horror when I open my mouth. I no longer need perfection. This is the level of quality I desire.

We need to evaluate many areas of our businesses and assure ourselves we are not overpaying for a level of quality we don't need. Or, conversely, producing a level of quality our customers don't value.

Over the past few years, the manufacturing industry has undergone a substantial digital transition. By using the most recent developments in robotics, the Internet of Things (IoT) sensors, artificial intelligence, and communication technologies, Industry 4.0 enables smart factories that can operate autonomously. To reduce product development time and improve product quality in the face of Industry 4.0 disruption brought on by a rise in product complexity, companies that develop configurable products and manufacturers of products with many different variants must create a single collaborative source of configuration truth. Let’s consider the factors that go into accomplishing this goal. 

IoT isn’t really about the internet. It’s not even about things. It’s about connections. Think of an IoT system as a series of data-sharing events enabled by devices—along with computing power, software, and connectivity technologies that round out the ecosystem. Data flows to devices for over-the-air (OTA) control and software updates. At the same time, data flows from devices to other devices, digital systems, or user platforms. 

There are many different conceptions of the Internet of Things (IoT), and varied viewpoints about how to use it. But these are finally coalescing into an understanding of the IoT as an ever-growing collection of digital connections among and between physical "things."

The Internet of Things, a term once applied to almost any “smart” gadget connected to the Internet, is becoming more useful, more complex, and more of a security risk as the value of data continues to grow and more people depend on IoT technology.