Not many technologies have been embraced as quickly as analytics. Analytics has moved from helping us gain insights into past events, to real-time monitoring using dashboards, and now to predictive analytics that can forecast what can happen hours, days, weeks and months from now.
Many industrial environments are becoming so complex to manage that making decisions without analytics may jeopardize millions of dollars’ worth of products and services. Executives may continue to make gut-feel decisions, but taking into account the risk this might create, they’d also want to delegate some of the decision-making to analytics. Data and analytics are becoming common in health care, energy production and manufacturing. As a result, patient outcomes are improving, power generation is becoming more efficient, and manufacturers are improving product yield and quality.
As machine learning and predictive analytics become more widely used in commercial settings, new trends are emerging. Predictive analytics is now accessible to people not traditionally engaged with analytics, such as those who manage shop floors and shipping docks. This is giving rise to operational analytics, which answers questions like, “What is actually happening on my production line now and what does this mean for us in the near future?” This new segment of analytics users is trying to figure out how to work smarter and act proactively.
Translating insights into actions and value
So, what can industrial users of analytics do to more quickly see value from their investments in analytics? When we begin working with a new customer, we immediately determine what the high-value question is that they want to answer. For example, how to extract more oil from a certain location? Or how to care for more patients in a hospital with the same staffing levels? If they have the data to help answer those questions, we then need to get busy building the right analytics.
We’re at a very nascent stage of analytics adoptions. And companies are willing to adopt analytics only after their peers or vendors can demonstrate results. This growing body of results will lead over the next two to three years to the broad expansion of industrial predictive analytics.
If companies are convinced with the results demonstrated at this stage and decide to make an investment in predictive analytics, they’d also want to turn insights into action. The best place to start is with a persona and decision inventory — who is making the decisions, what are the types of decisions you’re currently making and which of these are the truly high-value ones. As you work through this auditing process, you’ll identify the best places to focus your analytics efforts.
It’s also important to avoid a big bang approach to analytics. For example, instead of spending a lot of time in the data gathering before proof of concept, bits of data could be pulled out for use in rapid prototyping. This data could be read to identify if there is value and to check whether more data has to be pulled for more value. Starting small and prototyping your way to the right data sets help you find high-value answers and quickly deliver value to the organization. The secret of getting ahead, after all, is getting started.
Legal disclaimer: The views expressed are those of the author only and do not represent the views of any of the member firms of Ernst & Young Global Limited.