The Case for Reorganizing IT for Hybrid Environments

Could the integration challenges of cloud computing trigger an organizational shift for IT?

IT tends to swing from centralized control to decentralized control every two or three decades, anyway. In times of strong centralized control, which is what we’ve seen in recent times, most of IT will report to a CIO. When the pendulum swings to de-centralized control, you’ll typically find IT workers reporting to specific LOB managers, although a smaller central IT organization may remain.

Already, cloud and SaaS have moved many IT decisions out of the CIO’s domain and down into the line of business.  In fact, Gartner is predicting that chief marketing officers alone will outspend CIOs on technology by 2017, according to “Maintaining IT Relevance in a Hybrid Environment.”

 In that Cloud Times column, Scribe Software’s VP of Technical Resources Mark Walker discusses the unique challenges of managing IT in a decentralized, cloud-based environment.

After reading it, I’m starting to think that the integration challenges of cloud may be unique enough to swing the pendulum toward decentralized IT, at least for the foreseeable future.

“This is a transitional time for IT teams,” Walker writes. “The cloud has decentralized infrastructures, but a global perspective of corporate data is still critical to preserving a shared knowledge base. In this world, IT managers must seize the opportunity to lead by taking the role of business advisor rather than technology practitioner.”

In the past, IT has dealt with this by fighting back — regaining control of technology spending either by policy, budget or political means. Centralized data and the costs of integration work provided a strong justification for giving CIOs dominion over line of business applications and systems.

CIOs could fight back, but as we’ve already seen with cloud, that’s a losing battle. Instead, Walker recommends a more tempered, cooperative approach:

“… the CIO must rethink IT’s role to become part of the solution by using carefully planned and integrated cloud apps to empower LOB managers for an improved bottom line,” he writes.

Might it be time for enterprises to consider a decentralized approach to the IT organization and integration work?

I certainly don’t know, but it strikes me as just crazy enough to work, IF a few practices are in place, such as:

• An Integration Competency Center or perhaps a Data Center of Excellence that could serve as a central governing body for data rules, data quality and governance guidelines, integration requirements and best practices
• Perhaps consolidating on one data management platform or at least a specific list of cloud and on-premise integration options
• Solid API management
• A smaller centralized IT division that worries less about LOB technology, and more about strategic use of data or other strategic technology initiatives

Advantages could definitely come from such an approach.

First, it would free central IT and the CIO from tasks like enforcing minor technology issues and maintaining and integrating every application and system in the organization.

Second, decentralizing IT could make it easier for IT to push data quality and governance further into the business.

Third, it might make strategic integration a reality, freeing CIOs to focus on the big picture while putting IT workers on the front line, where they raise integration questions while the line-of-business manager is still evaluating solutions, rather than after they’re implemented.

Finally, it may actually better support cloud solutions, since integration is one of the major reasons for abandoning cloud projects.

Regardless of how you structure the IT organization, Walker points out that you have to address the data silos forming in hybrid environments. He offers a few suggestions on how to do that, including:

• Adopting an agile approach to data integration projects; specifically, this means prioritizing and iterating the work.
• Using integration platforms that support repeatable integration processes (SOA, anyone?).
• Leveraging data mapping templates as a starting point to cut down on time and effort.