Prescriptive analytics is a bit of a unicorn – a thing of beauty, but rarely seen. I think that’s about to change, with prescriptive analytics and the Industrial Internet of Things (IIoT) enjoying their teenage years together. Other styles of analytics (Describe, Discover andPredict) are inherently dependent on subject matter experts (SMEs) for interpretation. That is, to a large extent current analytics tools and applications present data and information, but with little related business context. So, a SME is required to infer the context and work their way towards a decision, based partly on data, partly on their expertise, and partly on intuition.
Prescriptive analytics will start to change that. Prescriptive analytics will provide the business or operational context that’s often been missing from other styles of analytics. While that’s a good thing for analytics in general, it will be critical for the Industrial Internet of Things. As I’ve noted before, IIoT data will check all the boxes for big data – it will come from new sources, it will arrive more rapidly than before, and it will accumulate into massive volumes. And to top it all, weighty decisions may need to be made quickly in order to optimize production or minimize the downtime of an expensive asset, whether onshore, or offshore.
Some of those decisions may be automated, but many will require some degree of human assessment before action is taken. In critical, time-sensitive situations, context is…critical. Here’s a simple example. Imagine you’re late for your flight, rushing from a customer meeting to the airport. You leave the clients office, decide to cross the street and walk a block to a hotel to get a cab. (OK, so you’re smarter than me and you’ll already have used Uber or Lyft, but just let your imagination run for a moment, it’s good for you I promise…). You walk out the door, you’re about to cross the road when you see a flashing stop sign! Decision time: Wait, or go for it…? Hard to say, because you just don’t have enough context to make a quick – and more importantly, deadly accurate – assessment. But, what if you saw the countdown timer – together with a rapid assessment of your athletic ability – help you make an instant decision. And that’s a very simple example of context – additional, relevant information that’s needed to make a fully informed decision. You don’t need to be a subject matter expert on the traffic patterns at that particular junction, the technology takes care of that for you.
So as industrial data becomes more varied, moves faster and accumulates, decisions will become more complex. Making complex decisions quickly and accurately requires help from technologies such as predictive analytics and machine learning. But, with complex data fueling complex decisions, the insights surfaced may not be trusted immediately. Actions recommended by algorithms may only make sense in the light of context. For example, it may not be enough for a computer to state that a piece of equipment is going to fail. An operator also needs to know why that alert has been made – for example highlighting the fact that similar equipment failed under similar circumstances 3 times in the past year. And that’s why prescriptive analytics will grow with IIoT.
About the Author
David is a Senior Analyst at ARC Advisory group, researching the use of business analytics across the industrial enterprise. David uses his 20 years of experience from many industries to research the art and science of getting the right information, to the right people, at the right time. Choosing an appropriate business intelligence (BI) or analytics solution is vital. But, many other factors are also crucial for a BI project to create lasting business value. With this in mind, David’s research has two goals: First, to help technology buyers to get the most value from their investments in BI and analytics. Second, to help suppliers shape their analytics product and marketing strategies. Immediately before joining ARC, David researched business analytics for the Aberdeen Group, serving clients such as SAP, IBM, Qliktech, and Tableau. Prior to that, he worked in marketing roles for companies such as Oracle, Cognos, Dimensional Insight, and Progress Software. David’s education includes a BS in computer science from the University of Hertfordshire, and an MBA from Cranfield University, both in the UK. LinkedIn: http://www.linkedin.com/in/djowhite/ Twitter: @addicted2data