You can’t do it all. You can’t know it all.

That is both refreshing to admit and frustrating to live with.

As a CIO, it is unlikely that your educational path included much of the data analytics explosion of which you are now a part, and in some respects, expected to lead. You can’t do it all. You can’t know it all. But you better know what to do with it.

It’s surprising how once established in a profession, leaders discover disciplines emerging that are a foreign language to their common vocabulary. Research and statistics? Maybe you excelled in them, but most CIOs who climbed their way to the here and now managed to avoid them.

But it is all here now. Analytics is challenging business structures, and it is fueling rapidly the need for new and different talent.

In the end, it is a business function, and more CEOs and stakeholders are acknowledging its value.

The question CIOs raise with me is how to lead it, specifically, whether it is centralized in IT or spread across business units. Centers of Excellence had a run, but it hasn’t translated in practice as well as it has in design. But we are learning.

Centers of Excellence can result in greater efficiencies, better quality of product, reductions of cost, consistency and risk-mitigation of knowledge management. However, they are not always aligned with the unique needs of business units and can result in bottlenecks and backed-up ques.

It is not an either/or proposition. The question is not whether data analytics should be centralized, but when it should be.

The objective is always to get actionable insights to decision makers, and for business units, it’s not about the data you need but the decisions you make.

That being so, the role of the CIO and IT is broader. You are the architect of learning communities.

Essentially, you need a person who has experience and expertise in data analytics building teams who can be trained. It is a skill set that can be learned. And yes, you want to spread out data analytics across business units, because each of those centers will be able to collaborate and strengthen a company-wide culture and practice of data analytics.

But until such teams are built, data analytics may be an IT function. Remember, build the environment before creating living things. It’s the way of life. If you want something to sustain, address the environment first.

The revolution of data analytics has caused IT leaders to pump out “living things” without a path for sustainability.

Reframe your role. You are not the chief over data analytics, even in delegated fashion. You are the architect of learning communities that serve and prosper the decision makers across all business units.

You can’t know everything. You can’t do everything. But today, to compete, you must know and do that.