11: Organizing for Success
BI is typically implemented within a department of business unit, generally when a unit is underperforming and management on that level is desperate to find a solution (as an underperforming unit generally gets negative attention, the kind that results in layoffs, reorganizations, or even outsourcing of the unit's function). However, the goals and requirements of a department may be inconsistent with the goals of the enterprise. This may result in a system that focuses on the business unit's needs and uses a quick-and-dirty approach to achieve short-term success.
The author notes that there are some instances in which BI can be implemented on the department level rather than the enterprise: IF a department is self-contained, does not work in tandem with other business units or leveraged shared resources, and when the success of the unit is independent of the enterprise performance objectives. (In other words, when the "unit" is an independent line of business.)
Otherwise, BI should be considered at an enterprise level. It should focus on the goals of the entire company, provide resources that can be shared among the various departments and units, and have more of a long-range perspective for attaining a sustainable advantage.
In some instances BI has been "pioneered" in one department and later attempted to be rolled out on a more widespread basis. However, this approach has not been successful, in that the solution does not scale well, or must encompass diverse and sometimes contradictory requirements. And so, whatever is built at a department level generally must be scrapped and an enterprise solution built from scratch.
This is not to say that an enterprise solution should focus solely on the grand scale and ignore the needs of business units, but to focus on elements that are common to all departments in a way that it can be adapted to provide for the specific needs of a unit. The author refers to the "tragedy of the commons," in that business units will act in their own self-interest regardless of the impact to the common good, which points to a need to have BI managed at the highest level, in order to ensure that those who control the funding for projects take into account the needs of the organization as a whole.
A few anecdotes are provided of instances in which a business unit that "owned" a BI solution (or even the data to be fed to it) refused to consider the needs of other units in developing system requirements, and even refused to provide access to the data because they felt a sense of ownership of it.
Organizational Support
An effective approach to managing BI is to develop a "steering committee" that includes senior representatives from the various business units that will support and utilize the enterprise BU systems. This committee has strategic oversight of all BI initiatives and the authority to resolve issues and conflicts among business units.
Additionally, a BI "competency center" or "center of excellence" can be organized for promoting the effective use of BI within an organization. This should be a cross-functional team who supports and promotes the use of BI throughout the organization. This is differentiated from a project team, who has a clear scope, set of deliverables, and timeline, as the competency center takes a broader, more long-range view and does not have specific short-term goals.
The author provides a list of "guiding principles" for those who have oversight of an organization's BI systems:
- BI should be viewed as a strategic asset with a goal of competitive advantage
- BI should be considered on the enterprise level, and any initiative should be considered from that perspective
- The IT group should support and deliver solutions that are driven by priorities established by the business (specifically, with a goal of delivering business value rather than gimmickry)
- The company should seek to borrow "great ideas" from other firms rather than cleaving to a "not invented here" attitude (EN: this would seem at odds with competitive advantage - imitation and innovation are opposing forces, though it may be more productive to seek a middle ground between the extremes)
- Where data is concerned, errors should be addressed at the source rather than by creating a patch
- The success of the BI system should be measured in terms of its impact on the business
- Any service or database that can be "shared" will be centralized, including the hardware, software, and systems governance, to prevent the balkanization of resources.
- While the economics of buy-versus-build may influence decisions, the long-term interest of the business should drive to obtaining solutions that are adaptable, rather than limited.
- The BI systems should seek to be on the leading edge, but stay short of the "bleeding edge"
In addition to governance, the cost of maintenance and support should be centralized: if BI is to serve the enterprise, it must be funded at the enterprise level, which includes the maintenance of the systems and the training and support provided to employees. However, in instances where businesses have specific BI goals, it is reasonable to bill the cost of their projects to the business unit that initiates them.
Case Study: Corporate Express
The author describes an initiative at Corporate express to centralizing the BI resources within the company. As with most firms, the company had many isolated "decision support systems" for its various units that stored data and provided reports, which meant that information was bottled up in separate departments and the tasks of gathering and analyzing the data were highly inefficient.
The problem came to a head in factual disagreements among business leaders, who were constantly arguing over the validity of the "numbers" rather than discussing business priorities. This resulted in the centralization of DSS positions, reducing 31 "regional" roles into a group of four central experts, and placing authority over them with the finance group rather than IT, which was relegated to a supporting role.
When creating the central unit, the company sought individuals who would relate to the business first, and to see the information systems as a means to an end, gave them a clear mission, and provided support from a senior executive level.
BI Team Leaders as Level 5 Leaders
The author refers to Collins, an author who examines organizational leadership, and defines various levels of leadership. His "level five" leader is a professional who focuses not on individual performance, nor on goals that pertain to a business unit, but who seeks to achieve goals that serve the entire company. This characteristic is appropriate and desirable in choosing directors who will be given authority over BU within a company.