Worried About Data Collection? You Should be More Concerned About Data Contextualization
Data Contextualization trends are evolving rapidly. Do you know the accurate cost of a single quality rejects at every stage of your manufacturing process?
Do you know what impact an unexpected process deviation early in the manufacturing workflow will have on the quality of the final output? Do you understand how quickly you need to respond to every downtime incident before it begins to impact your supply chain (and consequently, your P&L)?
If you’re the average manufacturer, you probably answered “no” to at least two (if not all three) of the questions above. And that is not to say the average manufacturer is not collecting data from their manufacturing floors to guide decision making.
Every company can collect data. That part isn’t hard to do.
But it’s what is done with the data once it’s collected that separates the pretenders from the contenders.
With all the talk around digitization, data-driven decision making, and Industry 4.0 initiatives, we’re seeing an uptick in manufacturers beginning to invest in data collection programs to test the waters. With sensors becoming cheaper and the tools to collect and disseminate information easier to access, suddenly, a lot more data is being generated on manufacturing floors.
This is exciting, but don’t let this effort overshadow the bigger purpose. In fact, I posit that data collection, while revolutionary, is still only table stakes. What will set the winners apart will be those that can contextualize the data to support better decision making across the enterprise – shop floor to top floor and supply chain to the customer.
Consider a very simple example from the world of maintenance.
Capturing temperature data from motor bearing offers insights into the health of key components of that motor, with the rising temp being an indicator of reduced bearing performance and impending failure. But contextualizing temp data with spindle speed can offer further insight into why the temp may be rising (higher temps at higher speeds is normal). Add additional data from 3-axis vibration measurements, lubricant quality, maintenance history, and electrical power draw, and now you have a full contextual view of the health of the motor along with deep insights to help diagnose and address non-ideal behavior.
What if you could further contextualize this data against 25 similar motors across your facility to detect anomalies and accelerate learnings?
The likelihood of uncovering high-quality and dependable non-obvious insights (which will ultimately improve decision making and business performance)is suddenly within reach.
Additional layers of data help create a rich multi-dimensional context that a uni-dimensional view doesn’t offer. The more dimensions you have, the more relationships you see, and the more context you can create—ultimately delivering richer, more meaningful insights.
A manufacturing operation, simply put, is a significantly more complex version of the motor example above. The moment raw material is received, every single action impacts the quality and timeliness of the output from your facility. Equipment maintenance impacts quality. Quality impacts supply chains and customer satisfaction. Supply chains and customer satisfaction impacts P&L.
Timely responses to maintenance work orders, accurate tracking of quality checksheets, costing of individual parts and the impact that should have on financial decision making, employee satisfaction correlated with efficiency metrics, etc. all impact one another throughout the process.
The number of relationships between data dimensions increases exponentially as the system complexity (and consequently, the number of dimensions) increases. The ability to map that complexity and make informed contextual decisions is the biggest challenge that manufacturers face in their Industry 4.0 journeys.
The truth us, the ripple effect of a single action, delay, error, etc. is already modeled in other industries. Consider the complex models that drive the airline industry – where regardless of carrier, aircraft, origin, or destination, a single delay can significantly affect the entire network. The airline industry models in additional complexity to study the impact that external factors like weather have on the network.
It’s clear that this is important, and there are broad implications. So, how do managers start to gather contextual insights and model these network effects within their enterprises?
It begins with creating a single source of truth for all manufacturing data. When data exists across disparate IT systems, developing contextual relationships becomes tougher.
One manufacturer we at Plex work with, Argent International, was able to do just this. When they tracked machine uptime and maintenance activity independently, with the use of a single source of truth and correlated insights, they were able to improve their notification processes and delivered 20% sales (capacity) growth with fewer employees.
Data collection is table stakes. Data correlation and contextualization is where value is created, and the right technology partners can help manufacturers get there faster.
As we progress further into Industry 4.0, smart factories will be defined not only by their ability to connect to machines but also how they leverage the contextualization they have at their disposal.