By Keith Staninger, digital solutions and controls portfolio leader, Ingersoll Rand

After years of working with customers to implement software and systems that gather data, I’ve noticed the process is often implemented backward.

I was recently at a plant that had spent a significant amount of money collecting data into a plantwide supervisory control and data acquisition system (SCADA). That’s a fancy term for a place to keep all of your data. Someone had told plant leadership that the cure to the plant’s problems was having all data in one place – the idea being that if all of the data was in a software system, plant personnel could make decisions more easily.

As we conducted the walkthrough, I was impressed by the amount of data they had collected! As we finished, we admired the output on the screens that showed plant data from multiple systems. The maintenance manager quietly said: "It looks much prettier now. We still don't know what it means to us, but it’s much prettier."

Data exists to solve problems. The problems have to be identified before the data is collected for the data to be able to add value. I see a lot of customers like this one who start gathering data while completely foregoing goal-setting. When desired outcomes aren’t established, no one knows how to apply the data once it comes in it. Gathering data alone doesn’t solve the problem.

In decades past, every plant had a seasoned employee who had the skill to analyze metrics and understand deeper implications. Now, with the loss of the knowledge worker in manufacturing plants, it’s not a lack of data that stops companies from improving efficiency and saving energy, it’s the gap of knowing what to do with the mass of data generated.   

My advice - don't be afraid to keep your data in the system it was born in. You may not be able to get the ROI from a large software implementation if you don’t have an expert on hand to curate and scrutinize the data.

Instead, hold the OEM, a vendor, or the engineering team accountable for reading and analyzing the data in real time. Listen for THEM to tell you what the data is saying. Better yet, look to IoT-enabled equipment with machine-learning mechanisms that will tell you what the data is saying.

A notification that tells you that you may have a clogged filter? That’s helpful and actionable.

A notification that tells you that you that your differential pressure is high? That’s helpful!

A notification that tells you to change your filter? Now we’re talking!