-
-
Marketplace
-
Channel Resources
Articles from this Site
-
One brand, one Web site! DM Review is now the home of all the content you're used to at BIReview.com and much more. If you are registered at BIReview.com, you're already registered at DM Review. If not, take just a moment to sign up for all the free services we have for you at the new DMReview.com.
Enterprise-Level BI
High-quality information has become a recognized asset that no enterprise can do without. Nonetheless, many - if not most - organizations still suffer from an inability to exploit information in a satisfactorily manner. To overcome this, business intelligence (BI) solutions are implemented, aimed at facilitating the access and efficient use of an organization's information. Having gone through various BI implementations over a period of 10 to 15 years, many organizations are now faced with the need to stitch together disparate and departmentalized BI solutions, methods and approaches. The goal is a coherent BI output (reports and analyses) on an enterprise level, thus allowing a more complete understanding of the state of key activities. This is quite a tall order.
There are a number of obstacles that hinder efficient enterprise-level BI. One major obstacle concerns organizational structure. Even though many organizations are recommended to keep reasonable decentralization and a departmentalized independence in order to be reactive to ever-changing market conditions, this decentralization does not work well when trying to execute enterprise BI. Simply put, the organization runs into the "too many chiefs" syndrome. Too many people have a say on how the information infrastructure should be implemented with all its data marts, reports, OLAP cubes and other components. Given the great variety of demands and needs, complexity becomes a major obstacle to efficient enterprise-level BI. The result far too often is that the enterprise-level BI implementation becomes slow and tedious.
Some of the common organizational obstacles to enterprise BI are:
- Departmentalized budgets, which require discussion and consensus on who shall pay for the centralized enterprise BI.
- A departmentalized organization often means that key personnel are attached to certain departments and not available at the corporate level. These key people include project managers, solution experts and solution architects. In practice, it can be difficult to get hold of these people when launching an enterprise level BI initiative.
- A departmentalized organization also implies that business users risk being stuck in their own line of work without clearly comprehending the information dependency across departments and processes.
Thus, achieving enterprise BI is often hindered by organizational, rather than technical obstacles, abetted by a lack of high level know-how on what's needed to actually implement enterprise BI, even though this can be temporarily bridged with external service providers and training.
THE BI COMPETENCY CENTER
The most-contemplated answer to the problem of enterprise BI is the establishment of a Business Intelligence Competency Center (BICC). The purpose of the BICC is to leverage the organization's full information management capability so it can optimize decision support at all levels. Its overall vision is to secure information availability and timeliness for business use and to support business in exploiting the intelligence in information. Thus, the key driver for implementing a BICC is to better use the organization's existing but often departmentalized and unconnected information.
The most evident starting point for the BICC is to address the information landscape instead of the system landscape. How the information landscape is architected will in turn bring about efficient selection of systems to address the system landscape. The main points to take into account when forming the BICC are:
- Fix the source, not the messenger: A balanced focus on the quality of underlying data, the data storage (often a data warehouse or data marts) and the accessibility of the user interface (reporting tools) is essential to leverage the organization's information potential. If there is too much focus on the user interface at the expense of the other parts, the BI output becomes limited in its use, no matter who tries to create actionable information from it.
- Work at the forefront: More work that is taken into account early on in the requirement's process means that less that will go overboard later because it was "forgotten." At the same time, it is essential to keep a pragmatic approach where results matter and that the organization does not get bogged down in trying to plan for every eventuality. This kind of overly-detailed planning is simply not possible.
- Process integration is important: Make sure that business processes, their detailed business events and their interactions are understood. End-to-end BI management takes into account the information created when such business events take place. All these events need to be collected and then structured into defined information flows. If not, the resulting information flow will be only vaguely understood, resulting in information that is poorly adapted to support the business.
- Governance is a requirement for optimal information management: Responsibilities cannot be shared between different departments unless there is someone who can make the final calls. Therefore, a BI council consisting of both business users and IT, with a top-level sponsor, is a minimum in order for the BICC to succeed. Truly holistic BI means that business and IT must evolve to a new level of collaboration.
- Clarify why the BICC is established: Clearly communicate the objectives of the BICC, including a vision, roadmap, blueprints and governance, as these are key elements to guide an integrated BI environment. This all comes down to aligning key elements such as data quality activities, reporting set-up, and overall enterprise information organization, into a coherent BI strategy.
Define how the success of the BICC is to be measured: Success factors you may wish to consider:
- A high degree of reports stemming directly from the BICC compared to the number of reports built by business users themselves;
- A high number of business users using the BICC facilities compared to the total number of concerned business users;
- A lower annualized costs of producing business applicable BI output, leading to lower cost of ownership for the BI related tools and services;
- The number of reports and analyses in relation to the number of business users (where many reports for a smaller number of users should not necessarily be considered as something good or optimized);
- The reduction in the number of tools and service providers, allowing for more standardization and therefore easier BI development and maintenance.
The good thing about implementing the BICC is that established methods can be used. The business process reengineering approach can be used; functional blocks or component business models can be referenced; and existing BI tools can often be further leveraged. Given the strategic importance of a well-working BICC, it might also be a good idea to reflect on how much is handled internally and how much is in the hands of external suppliers. Even if the actual data and reports and analyses may be prepared by a supplier, the governance needs to be firmly in the hands of the organization using BI. It is indispensable that the organization takes an information centric approach to its business. This information-centric approach starts with the organizational chart so that the organization's intents and operations are clear to the business users.
GOVERNING THE BUSINESS INTELLIGENCE COMPETENCY CENTER
For a centralized cross-business approach for the BICC to succeed, the organizational chart must be taken into account. This will facilitate existing organizational structures and business culture and ensure that the BICC organization is not deviating unnecessarily from the organization's common way of doing things. A "middle way" is to keep a somewhat distinct division between IT and business on an operational level with clearly defined roles, but with a cross-business BI council governing activities and being accountable for the BICC's success. A high-level model for the BICC organization is found in Figure 1. In this chart, there is a clear distinction between information and data, where information is business driven and defined as organized and actionable data. The data part remains IT driven, thereby adhering to common practices in most organizations.
REFLECTIONS ON THE BICC
For a common framework of needs, solutions, and approaches to succeed, it is essential that BI strategy goes hand in hand with business process strategy. This will drive business process integration while allowing a comprehensible information flow that can be properly exploited. A BICC with its governing BI Council will effect the bringing together of business processes, thereby facilitating process integration and common information sharing.
Even though only 20 to 30 percent of large and medium-sized organizations already have a formal BICC (according to Gartner Inc.), there appears to be a clear trend in this direction. The common challenge is the organizational and cultural issues that must be overcome, and realization that centralization of business activities typically bring a negative response. It is important to quickly give something back to the business users in the form of reports and analyses with additional facts and figures that the business can act upon.
In the end, the ideal BICC will serve as an information "One-Stop Shop," the one place from which an organization's BI activities are centralized and controlled, allowing for best possible quality, and consequently, confidence in the information.
Rachel Chew is a solution expert in business intelligence at Swedish telecom solution provider Ericsson. The ideas expressed here are based on personal experience and do not necessarily reflect the situation at any particular company. She can be reached at rachel@jermyn.net.
Gabriel Fuchs is a senior consultant, author and BI expert. His thoughts are also based on personal experience and do not necessarily reflect the situation at any particular company. He can be reached at sgfuchs@bluewin.ch.
For more information on related topics, visit the following channels:
