I doubt there’s a manufacturer in the industrialized world who hasn’t heard about the Internet of Things and how it’s going to drive the next industrial revolution. And it certainly could. But whether the IoT actually delivers on its promised value has little to do with how well the various sensors collect and transmit data; it has everything to do with how we use that data. And it starts with deciding which data to use.
Machinery that is equipped with sensors and connected to the IoT generates mountains of data. One GE wind turbine, for example, is equipped with 20,000 sensors that generate 400 data points per second. That works out to more than one million data points per hour – just for one piece of machinery. If you think about that happening all day long, all across your factory floor, it’s easy to see how the sheer volume of data being collected could render the whole thing useless if you’re not careful. Nobody can work with and act on that much data, and a lot of people simply don’t know where to start. But it’s critically important to give careful thought to which data you’re going to pay attention to. To get the most benefit out of the IoT, ignore any data (for now, at least) that doesn’t meet these three criteria: relevant, actionable, and strategic.
If you took logic or statistics courses in college, you might remember having discussions about “correlation vs. causation”. In a nutshell, it means that just because two data points are related, that doesn’t mean either one causes the other. Here are a couple of examples: The divorce rate in Maine has a 99.26% correlation with per capita consumption of margarine. And there’s a 94.7% correlation between per capita cheese consumption and the number of people who die by getting tangled in their bed sheets.
Obviously, it would be absurd to suggest that there’s a causative relationship in either of those scenarios – despite the fact that, when graphed, the trends match each other almost perfectly. Your factory floor isn’t so different. Sometimes you’ll see a pattern in your data, but that’s all it is – a pattern. There’s no causative relationship there; it may be interesting, but it’s not relevant. It takes an in-depth knowledge of your business, products, and processes to tell the difference – and a whole lot of discipline to set aside the information that doesn’t matter.
Not all relevant information is actionable. Once you’ve determined that a particular data set is relevant, next you have to decide what – if anything – you can do anything about it.
Let’s say you’ve identified a causative relationship between the outside temperature and your equipment failure rate. Can you do anything about the outside temperature? Of course not. But you could use that information to fine-tune your preventive maintenance processes, performing more frequent inspections and replacing parts sooner during hot weather so that your business doesn’t suffer unnecessary downtime. It’s actionable – but you’ll only recognize that if you have a thorough understanding of your plant’s processes.
Now you’ve narrowed things down to information that’s both relevant and actionable – but there’s still one more step to go. You have to ask yourself how it fits into the business’s strategic plan. As an example, let’s look at a manufacturer of defense equipment. At this particular plant, they’ve figured out the optimal number of times to turn a particular screw during the manufacturing process. And they know that failure rates go up when that number is exceeded. So any time a screw is turned too many times, the part is flagged as an anomaly so that it can be inspected. Sound like a minor detail? Maybe – but quality is a business imperative for this manufacturer. So the data is not only relevant and actionable – it’s prioritized right there in the strategic plan.
And then there’s the oil and gas industry. Experts in that industry know that a 1% increase in the efficiency of the pumps that are used to pull oil out of a well works out to more than half a million additional barrels of oil a day. At $100 per barrel, that’s more than $50 million each day. That’s a no-brainer, right? Not necessarily. If the business’s strategy is to focus all of their resources on identifying and developing alternative energy sources, increasing productivity – even to the tune of $50 million per day – might not belong on the to-do list. Sure, it’s a lot of money, but it doesn’t support the company’s strategy deployment. They might decide to come back to it at a later time, but, for now, the smart decision would be to stick to things that are in line with their focus on alternative sources of energy.
According to a white paper published by PwC and the Manufacturing Institute, the manufacturing industry stands to gain about $4 trillion in value (over the next decade) from the Internet of Things, primarily due to increased revenues and lower costs. But you’re only going to receive the full benefit of the IoT – and at a reasonable cost – if you focus on the things that matter. Don’t get distracted by the sudden influx of information. Sure, a lot of it was information you didn’t have two years ago. But did you want it two years ago? Did you ever even think about what you might do differently if you had that information? If not, set it aside for now. Focus on the data sets that are relevant, actionable, and strategic.