Throughout organizational history, the decision of how much authority to decentralize and how much must go through the corporate hierarchy has been under ongoing refinement and debate. Many organizations experience times where the company is ruled by an iron fist at headquarters, with much of the authority and direction coming from a few key executives. At other times, front-line or distributed employees are empowered to make key decisions with loose guidance and little status reporting from headquarters.The decision of whether or not to centralize applies not only to decision-making but also to systems and functions. Lately, the discussion has become quite trendy regarding customer intelligence within businesses that work in separate business units, product lines, geographies or marketing units. Hewlett-Packard is undergoing this transformation right now. In the past, the HP way was extremely decentralized with much of the control and decisioning at the business unit level. However, HP is currently consolidating their systems infrastructure, including data warehousing, from 28 data centers down to less than 10.

From a customer intelligence perspective, what are the key factors in making the decision to centralize the function and systems? What would cause an organization to be more decentralized than others?


Many organizations currently have a distributed customer intelligence data architecture and are contemplating centralizing. In almost all cases, a sense of urgency to bring the customer data warehouse to market was the underlying decision to decentralize. Collaborating with other business units on the vision, design and use of the data warehouse was perceived as difficult and time-consuming.

Probably more than the fear of the initial implementation was the future direction of the customer intelligence infrastructure. Different business units did not want to propose system enhancements to a committee or governance board. As urgent as the initial development was, future enhancements would be equally urgent. Collaboration again meant delays.

Some of these organizations have left their options open by implementing in a decentralized fashion but collaborating on design and putting governance procedures in place to keep the design in close proximity. If the organization decides to centralize in the future, similar designs will make the effort easier.

Technology Redundancy

One of the most powerful cases for centralization is the technology environment and support staff. If you draw out the technical and production support landscape, it is typically easy to see redundancy that could impact cost. For instance:

  • Each location must purchase its own software licenses, which could include reporting, analytics, campaign management, ETL (extract, transform and load), databases and customer information cleansing.
  • Each location must purchases database and application servers in order to support the software environment.
  • Each location will need its own resources to monitor the system as well as production support monitoring or production batch processing.
  • Enhancements need to be rolled out to many locations instead of a centralized environment.

Each of these areas implies a level of cost, and the lure of centralization is the lure of cost savings.

Beyond the technology are the customer intelligence functions themselves, most notably ad hoc reporting and analysis, predictive modeling and campaign management. Like the technology discussion, it is conceivable that these functions can be centralized and possibly save money. The factors that drive the decision are usually the following:

  • Are there best practices to be shared across geographies or business units that could increase results?
  • Are there cross-sell opportunities across the different divisions that centralized analysts may be able to identify?
  • Who owns the number? If the business units are responsible for the sales or revenue number, the executives in charge of those units do not want to give up control of the resources that execute revenue-generating campaigns and programs.

Organizations successful at centralization have found the balance between the efficiency of centralization without the loss of business unit control of strategic direction or the ability to execute emergency programs. Some organizations leverage the centralized structure for the routine activities - management reporting, established campaigns, re-evaluating models - and the localized structure for more ad hoc or challenging exercises.
Centralization seems like a lower cost of ownership and a possible way to offload routine or even commoditized activities. However, the technology infrastructure to support the centralized environment coupled with support functions, such as campaign management, can surface hidden costs. You'll need to assess the cost of a data conversion to a centralized environment and what specific functions would be organizationally accepted in a centralized fashion. Further discussions can ensue if centralization is related to outsourcing.

Larry Goldman is president of AmberLeaf, a customer intelligence consultancy. He can be reached at

This article originally appeared on DM Review.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access