Manufacturing lessons can come from unexpected places.

We recently had some trouble with the dishwasher in our Somerville offices. The machine was behaving erratically. Sometimes, dishes would emerge sparkling and warm to the touch. Other times, they’d still be coated in lunch.

Our trusty appliance was leaving us with a decent amount of rework.

Troubleshooting the causes for flagging performance got me thinking about how a dishwasher isn’t so different from it’s industrial cousins on the factory floor. At its core, a dishwasher is a mix of hardware and software managed by a human operator that efficiently performs a task in less time than it would take a human. (Okay, for the dishwasher we’re talking convenience over speed, but the point stands).

The reasons it was performing poorly wound up being many of the same behind poor industrial machine performance.

What we learned applies just as much to asset performance management as it does to our kitchen.

Diagnosing the Problem

On the shop floor, there are a huge number of reasons an asset might not perform as effectively as it could. A tool could be worn down. A motor might need maintenance. Spindle speed could be too fast or two slow. CNC paths might be inefficient. Ambient conditions could affect performance. Operators might not be loading the machine properly. Machines might not be set properly after a changeover. The list goes on.

Some of these inefficiencies can be traced back to the machine itself (bad tool, breaking motor), while others are issues of machine usage (operator error, poor changeover).

In our case, we suspected the machine itself wasn’t broken. After all, sometimes it worked perfectly.

So that pointed to “operator error.”  

To test this, we experimented with a few variables. We tested how we loaded the dishwasher (layout in the tray, plate density, number of plates and mugs per load), different cycle settings, as well as adding pre-wash liquid in addition to the standard detergent.

Our Completely Scientific test produced results after just a few iterations. The only variables that had any effect on the machine’s performance were those related to how we loaded it. Density mattered. Quantity mattered. How dirty the dishes were when they went in mattered. Cycle setting, pre-rinse, heat dry–each had a negligible impact in comparison to the way dishes were arranged inside.

So ultimately, our dishwasher wasn’t broken. We were, in fact, experiencing a classic case of operator error.

What We Learned

The fundamental lesson here is that machine performance issues aren’t necessarily issues with the machine. They’re issues with machine usage.

Ours wasn’t an isolated phenomenon. Rather, it was one that cuts to the heart of inefficiencies in the modern factory.

Recent research has shown that over 70% of manufacturing problems are human related. So that stubborn machine problem in your factory might not be a machine problem at all. There’s a good chance that it’s just not being used properly.

Causes of Failure in the Factory

Causes of Failure in the Factory

Source: IHS Markit, DOE “Human Performance Improvement Handbook”, Noria research

The problem many manufacturers face–even those with sophisticated machine monitoring setups–is that machine data needs to be interpreted in a broader context. Even the most connected, IoT-enabled factory is only giving a partial picture of performance if it can’t account for the human element.

As a writer in Quality Magazine recently phrased the problem,

“Now that we can easily gather lots of data from machines, we can monitor them closely and predict problems before they happen, right? Isn’t that the promise of Industry 4.0? Maybe. Machine data isn’t useful if you don’t have a good way for a human operator to take action based on the machine’s data and you have a way to verify that the process has been done correctly.”

Think of what this means from a KPI standpoint. If we were to calculate our dishwasher’s OEE, the number would be way off, because quality and throughput were both impacted by poor usage. Same for OPE, which accounts for humans and machines. Given the invisibility of operators in the process, we couldn’t understand our process from a holistic perspective.

So when you’re out on the shop floor trying to diagnose a performance issue, think about what we learned from our dishwasher. Any asset performance management strategy needs to account for operators as well as machines.

Tulip gives manufacturers full visibility into human and machine performance on the shop floor. If you’re curious how Tulip can help with your improvements and challenges, get in touch to request a demo.