Predictive analytics are the most effective way to put big data to work, and cloud technologies both deliver new data sources and provide a scalable, pervasive platform for analytics. The intersection of these two technologies is creating new opportunities for value and new ways to exploit big data. It’s critical to understand how companies are thinking and how organizational thinking evolves as these markets grow in importance.

In 2011 Decision Management Solutions conducted a survey and associated research on the use of predictive analytics in the cloud. While it may not seem like that long ago, back then predictive analytics was just beginning to explode onto the scene. Now everyone is talking about big data, and the potential for cloud-based predictive analytic solutions is clear. It’s been an exciting 18 months so it’s time to find out what’s changed.

To learn how organizations are adopting predictive analytics in the cloud, we have launched a new survey targeted to business, IT and analytic professionals in conjunction with Information Management. (You can take the survey here.)

There’s been tremendous growth in the use of cloud infrastructure to enhance and expand the use of predictive analytics. This survey will provide insight as to where organizations are today, where they plan to go tomorrow, and how the perception of predictive analytics in the cloud has changed in the last 18 months. The results will yield valuable insights into emerging best practices, the kinds of technologies you should adopt at different stages, how fast you should move and what your competitors might be doing.

If you’re interested in some historical context, Key findings from the the 2011 study revealed:

  • It was all about the customer. Using predictive analytics in the cloud to improve customer acquisition and development was a big focus.
  • There was growing momentum. Over and over again respondents told us that they had bigger plans for the future, and thought those plans mattered to their organization.
  • Early adopters, especially those adopting decision management too, were getting an edge. Those already adopting predictive analytics in the cloud were more likely to have ambitious plans, less likely to have concerns, more likely to operationalize these analytics and more likely to take advantage of big data.
  • Data security and privacy were top concerns. But those with positive results tended to worry less about all the various concerns identified.
  • Industries varied. Financial services were most likely to have seen great results, telcos had ambitious plans and health care delivery was falling badly behind.

Full results of the 2011 study can be found here.

It will be interesting to see what’s changed and what trends we can discern. The different use cases and their adoption will be central to this. When we first looked into predictive analytics in the cloud we identified five use cases, but it’s clear now that we can simplify this down to three:

  • Pre-packaged, cloud-based decision-making systems, also called "Decisions as a Service." In this use case, pre-packaged, cloud-based solutions are purchased and used to provide decision-making based on predictive analytics. For example, packaged solutions offering next best action, marketing offer selection, fraud detection or instant credit decisions as a service.
  • Cloud-based solutions to define and build predictive analytic models. These solutions can handle data from both cloud and on-premise solutions. They add value by moving analytic modeling closer to the data available in the cloud and by taking advantage of the elastic nature of cloud solutions to efficiently support demanding analytic algorithms by assigning compute resources as necessary. For instance, building complex predictive analytic models in the cloud using many compute cores and data stored in a cloud-based system, uploaded from an on-premise solution or available from a cloud API.
  • Cloud-based deployment to embed predictive analytics. This use case is the use of cloud-based deployment to embed predictive analytics, no matter how developed, in existing systems that don't have their own predictive analytics. The target systems might be custom or packaged solutions and be delivered in the cloud or on-premise. For example, using cloud-based deployment to embed customer churn predictions in a cloud CRM solution or using cloud-based deployment to link internally developed propensity to buy models to multiple customer-facing systems.

Obviously, individual products often combine elements of multiple use cases, but it’s clear that these are the drivers of value for predictive analytics in the cloud. The changes in the adoption and perception of these use cases, as well as the role big data is playing, will be central to the research. This year’s survey asks about your experience with predictive analytics, what you think about the main use cases for predictive analytics in the cloud and also about drivers, obstacles and the role of big data.
I hope you’ll participate.

Author’s Note: Join us at, October 22 at 12:00 Eastern, to learn about the survey results (registration information will be available shortly).  We’ll also be publishing the results in a research report that will be available through the Information Management white paper library.