To Compete on Analytics, First Compete on Data!

competitive-adv-talent-mgmt-720x491I am seeing a major change in thinking in the analytics world as people realize that if they do not straighten out their data strategy, they will never really be competitive with their analytics strategy. Here are two quick examples:

Two weeks ago, I attended a Chief Data Officer (CDO) event. I expected it to be all about the role of the CDO and why anybody would take the job if they truly understood what the role was. When I got there I could hardly find a CDO in attendance. It turned out that the conference was actually an Analytics conference that had been running for several years. The reason for recent change to the CDO title for the event was that they had all come to the realization that if they did not have a strategy to manage the delivery of trusted and timely data, the rest of the effort was irrelevant or even possibly damaging to their organizations.

A while back I read “Competing on Analytics” by Tom Davenport. It was a great and very thought provoking book for its time (2007) and I got a lot out of it. The one slight negative thing I remember about the book was that while the conversation about analytics for competitive advantage was very stimulating, the concept of delivering data to support that strategy did not appear until very late in the book (around page 155 actually). Fast forward to a very recent airplane flight where I read his follow-on book “Analytics at Work.” In this book Tom Davenport introduces his DART framework. Guess what the “D” in DART stands for? Right, Data! In this book, the second chapter is about data and the coverage of the subject is very thorough. (BTW: I highly recommend reading this book.)

Here are two great examples of a major change in thinking that I am seeing more and more frequently. Sure, I work for a Data company and tend to talk to people with a data perspective, but I am seeing and hearing this everywhere now.

It is clear, from an architect’s point of view, that we are paying for decades of project-based architecture with little thought to how data might be a shared resource rather than locked into a specific project or application. The movement to cloud applications has only accelerated that trend with business groups within an organization standing up applications and analytics in literally minutes.

What should people do to compete on data? I would recommend several things:

Every other major function is managed with a standard set of tools that ensure repeatable processes, skill reuse, and a degree of automation. It is time to do the same for enterprise data management. It is time to standardize your data management tools. You may not be able to do this 100% without stifling innovation, but a recent survey of architects by Informatica showed that the thought-leaders were planning significantly more standardization than the average architect. It’s the way to deal with data volume and complexity while increasing IT efficiency. Any other strategy is highly risky at this point.

  • Design for data as a shared resource.All data onboarded should be prepped and managed in a way that ensures that the data is discoverable, usable, and manageable by any project that needs to use it, not just the current project.
  • Design for business self-service.You need to enable your business users to discover and use trusted and timely data by themselves, and without any IT assistance. IT is struggling to meet the need for speed of business value delivery demanded by the business. One way to free up resources is to enable business self-service. It will also dramatically shorten the time required for a lot of business-lead projects that require data. Think of all those QlikView and Tableau users you have. What if they could discover and manage the data for themselves?
  • Design for automation.Automation requires a standard platform for enterprise data management first. Next, it requires strong integrated metadata management across the platform. By collecting and understanding both technical and business metadata and matching that with actual user activity it is increasingly possible to create systems that automate routine data management tasks and provide intelligent recommendations for more complex tasks.

Let’s be clear. This is not just about tools. It is about Strategy, People, Processes, Technology (tools) and Metrics. In that exact order. If you don’t have a strategy and if that strategy is not aligned with the business strategy and goals, you will be going nowhere fast. The other items are equally important, but that will be another future blog.

For next-generation thinking on data architectures and management see Think “Data First” to Drive Business Value.

For a free download Informatica Rev, a tool that enables non-technical users to prepare data click here.