Skip to content

Peer Into The Analytics Crystal Ball For 2015

On December 10, I helped facilitate the International Institute for Analytics webinar to announce our analytics predictions for 2015. I will provide additional commentary on several of the predictions that I am especially fond of in this New Year’s post.


We had a very good discussion on the call about how a CAO differs from a CDO and why an organization might look into putting both in place. We at IIA expect these roles to continue to proliferate in 2015. However, I have personally seen a lot of confusion with respect to these two roles. Many people seem to consider them interchangeable or accidentally mix and match the requirements of the two roles together so that it is hard to tell what exactly they are looking for.

In the end, we laid out a few common points on the call that many of us agree help to differentiate these roles and also help to point out why both are important. After all, in a large company, handling all of the responsibilities I’m about to describe is more than one person can likely handle.

A Chief Analytics Officer usually will report to the business and have an analytics background. This role focuses on determining what analytics an organization should pursue, what methods will be utilized, and also what protocols will be put in place to accurately assess and measure the impact of those results. This person also will work to set the priorities for what gets done and will ensure that results are disseminated and acted upon as needed.

A Chief Data Officer usually will report to IT and will have a technology background. This role focuses on determining how to best collect, store, and make available the data required for analysis. The governance and security of the data is also a core part of this role, as is the maintenance of the infrastructure to support it. Clearly, the CDO and CAO must work together effectively and their success or failure will be linked.


I have written in the past about the need for successful analytics professionals to have good communication skills and an ability to tell a story to help business people understand analytic findings. Pamela Peel, CAO of the UPMC Insurance Division, predicted that story telling may even become a formal job. Within her organization, she actually hired a professional journalist to work with the analytics team to weave results into a more compelling story line. This was a very intriguing idea to me and one that larger analytics teams might do well to consider.


Ensemble methods have been around for some time. Recently, however, their popularity has been increasing even further. In my opinion, this isn’t just because ensemble methods work, but also is the result of a few other trends coming together.

  1. The capture and standardization of data is much more mature today, which makes it easier to get the data necessary for an analysis together. This, in turn, allows more time for analysis.

  2. A broad array of complex algorithms is now available and accessible. It is entirely possible for an organization to have many strong algorithms at its fingertips.

  3. The amount of processing power available at a cheap cost means that we no longer have to be stingy with our work. We can afford to execute many iterations of multiple algorithms today without costs spinning out of control.

  4. Software tools are starting to build ensemble capabilities in which makes it very easy to request and execute an ensemble process. Your organization should consider taking advantage of the “wisdom of the crowd” phenomenon and move into ensemble methods in 2015.


I have also written about various privacy issues in the past. One of the predictions that I hope does come to pass came from Jeremy TerBush of Wyndham. He predicts that privacy concerns will drive the creation of tools and services that allow consumers to take control of their data. Not only should you be able to decide if your data is shared and used, but you should be able to control under what circumstances and at what price.

If we can achieve this goal of giving consumers full control of their data, it will enable us all to determine what we’re willing to give up in return for benefits organizations can offer in return. We will all be empowered to make our own individual decisions instead of having others make those decisions for us.


The final IIA prediction I’d like to address states that analytics will continue to become increasingly embedded within business processes. I didn’t actually submit that prediction, but I fully agree with it. After all, my new book The Analytics Revolution focuses on this very topic. It is critical that your organization starts to move towards embedded analytics driving automated decisions. Otherwise, it is impossible to realize the full potential of analytics.

Good luck in 2015!