Grapevine, TEXAS — Analytics is evolving in new and potent ways, including the convergence of services and analytics and the continuing emergence of predictive and prescriptive tools, said speakers Tuesday at Gartners Data and Analytics Summit 2017.
In his session "To the Point: Convergence of Services and Analytics Is on Its Way—Take Advantage of It!", Jorgen Heizenberg, research director at Gartner, discussed the trend of service providers developing analytics offerings that can be supplemented by their services. Heizenberg noted that in 2013, Gartner predicted that in 2017, analytical applications offered by software vendors would be indistinguishable from analytical applications offered by service providers.
Using this prediction as a jumping-off point, Heizenberg foresees a shift from analytics simply being "in the business," to where analytics "is the business." Furthermore, he envisions this convergence creating a new demand-and-supply model that will leverage the opportunities of the ecosystem and innovation, as well as deal with the challenges of complexity and agility.
In this model, leveraging the ecosystem includes a demand for data and insights from other industries, and a supply of open data, curated data sets, and APIs, as well as micro services, containers, algorithms, and analytical applications. Innovation includes a demand for using analytics to stay ahead of the competition and a supply of acquisitions by consulting organizations, as well as software companies buying analytics companies. The challenge of complexity includes a demand for more complex analytics and a supply of algorithms, analytical solutions, and data science services. Finally, the challenge of agility includes a demand for faster insights and a supply of analytics embedded into operational processes, business applications, and platforms.
In his session “Embracing Predictive and Prescriptive Analytics,” Peter Krensky, senior research analyst at Gartner, outlined the four steps of analytics: (1) descriptive analytics, which identify what happened; (2) diagnostic analytics, which identify how it happened; (3) predictive analytics, which identify what will happen; and (4) prescriptive analytics, which identify what should be done.
Krensky further broke down predictive and prescriptive analytics into two sets of tools: predictive tools include the probability of a specific outcome, predicting a series of outcomes over time, and highlighting uncertainties, while prescriptive tools include predefined frameworks for choosing between alternatives and outcome-driven, constrain-based evaluations of interdependent sets of options. Put more simply, Krensky’s model places predictions, forecasting, and simulation under the umbrella of predictive analytics, while rules and optimization fall in the realm of prescriptive analytics.