January Issue

5 Things CIOs Should Know About Prescriptive Analytics

Eye looking at data and analytics
Credit: Thinkstock

First there was descriptive analytics, then predictive analytics. The next step is prescriptive analytics, which actually tell you the best action to take.


What's the followup to predictive analytics?  It's prescriptive analytics, which actually tells you the best action to take. Here are five things you need to know it.

1. It's the next step in analytics. First there was descriptive analytics, using data to describe current or past circumstances. Then came predictive analytics, analyzing data to predict a future outcome. Prescriptive analytics suggests the best option for handling that future scenario.

"Prescriptive tells you the best way to get to where you want to be," says Anne Robinson, director of supply chain analytics at Verizon Wireless and a past president of INFORMS, a society for analytics and operations research professionals. "If you want to differentiate yourself, the next step is the prescriptive tool box."

2. It requires more data integration. Data scientists typically spend about three-quarters of their time preparing data sets and only a quarter running analysis, says Forrester Research analyst Mike Gualtieri. That imbalance could worsen with prescriptive analytics, he says. CIOs can help by making it easier and faster to compile the required data.

3. The CIO role: Evangelist. Strategic-minded CIOs should study the technologies and then sell the opportunities. "CIOs need to become evangelists on how the business can bring improvement through prescriptive analytics," says Tom Davenport, a Babson College professor and author of Big Data at Work.

The organization needs to be strategic in its approach to prescriptive analytics--which requires collaboration among senior executives, says Rajeev Bukralia, an associate lecturer in the computer science department at the University of Wisconsin at Green Bay and the university's former CIO.

4. There's a need for speed. The results of a prescriptive analytics program could facilitate long-range planning, but they might also be needed to determine immediate actions in business processes. "There are some situations where we need to execute those prescriptive analytics in real time," Gualtieri says. Consequently, the IT infrastructure needs to be capable of analyzing information and delivering findings almost instantaneously.

5. It won't replace human judgment. "Prescriptive analytics isn't about technology," Bukralia says. Instead, it's about people asking the right questions and knowing how to react to the findings. So if a prescriptive analytics program is to be successful, the organization must have the right people in place and get buy-in from those affected. CIOs can help by lending their change management skills--and by knowing how to balance technology-driven information and human judgment. "You still need to give front-line workers some degree of autonomy," Davenport says.

Drexel and CIO.com announce Analytics 50 award winners
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies