Business Intelligence (BI) is becoming a focal point within many business organizations. Users require timely, accurate data to generate and distribute business insights—insights that confer a real, if only fleeting, competitive advantage to those organizations.  So, how does one go about building, supporting and operating a world-class business intelligence organization? 

Every executive and manager associated with a major business intelligence (BI) initiative wants to ensure his project achieves a meaningful return on investment and lowers the cost of ownership of the BI environment. However, it remains unclear just what impact the advent of new technologies such as RFID or the success of on-line commerce and new legislative initiatives such as HIPPA and Sarbanes-Oxley will exert on how companies contend with the tidal wave of new data and requirements. With the creation of data warehouses containing many terabytes of data now commonplace, older solutions used to manage smaller data volumes and address security issues no longer seem practical.

Gartner Group and other industry analysts have observed that BI continues to grow as a priority for large and midsized organizations. A growing number of companies recognize the value of adopting a strategic and broad-based deployment of BI solutions. This “vision of inclusiveness,” to use Gartner’s phrasing, has been embraced by customers and suppliers as well as internal business units. 

So how does one go about building, supporting and operating a world-class business intelligence organization?  In the following pages, we offer the knowledge and best practices the authors have acquired in the course of their work with leading business intelligence software providers. 

Business Intelligence: Definitions and Uses

Worldwide and within many industries, BI is becoming a focal point within many business organizations. Users require timely, accurate data to generate and distribute business insights—insights that confer a real, if only fleeting, competitive advantage to those organizations. Using analytical BI applications, users can collect, access, analyze, and distribute information in every area of the business. Recent improvements in technology have made it both quicker to retrieve large volumes of data and cheaper to securely dispense insights to all users who need it.

For the most part, corporations have employed different BI tools to implement different types of BI applications. Each technology was chosen because it supported one specific type of application extremely well—for example, simple reporting,  or Web-based interactive reporting, or on-line analytical processing (OLAP) featuring data ‘slicing and dicing’, or in-depth analysis through relational on-line analytical processing (ROLAP), or advanced statistical analytics or proactive information or alert delivery. Unfortunately, managing and maintaining this diverse set of technologies has become increasingly expensive and inefficient.

More and more, businesses recognize they can benefit from using a single, flexible, and architecturally-integrated BI platform that can:

• Support the full range of BI applications—reporting, analysis and information delivery

• Provide a powerful development environment and reports that are easy to maintain

• Access massive amounts of transactional data as easily as it supports workgroup datamarts

• Support and administer hundreds of thousands of users with high performance and little administrative burden

• Operate with centralized control commensurate with mission critical applications—fault tolerant, managed performance, secure

• Integrate business intelligence easily into other operational systems via  open APIs

 What differentiates BI tools from platforms?  According to the Gartner Group (“Magic Tools and Platform Quadrant”), platforms are centered on a modern technical architecture that includes extensibility to other applications. In much the same vein, Wayne Eckerson, director of research at The Data Warehousing Institute, has put forward twelve requirements for next-generation BI platforms. The implications of the above are clear: the marketplace demands a BI platform that can serve as the foundation for all analytical applications within the enterprise. One response to this has been the mergers and acquisitions over the past several years among several business intelligence providers, which represents an attempt to build out a more complete BI solutions offering. This tactic suggests a maturing of the business intelligence tools market. However, this approach still leaves users facing design, architectural and security incompatibilities that will have to be resolved over time. MicroStrategy 7i is most probably the first business intelligence platform that meets the stringent platform requirements identified by customers and industry analysts alike. Its advanced architecture has earned it a visionary position in Gartner’s Magic Quadrant for Business Intelligence Platforms.

 How to Achieve World-Class BI Applications

Structurally, organizations with successful BI applications are those that understand there must be a balance between meeting user requirements, ease of administration, and database design.

Over the past 15 years, five common ‘styles’ of BI have evolved.  These include:

•  Enterprise Reporting

•  Cube Analysis

• Ad Hoc Query Analysis

• Advanced Analysis

• Alerting and Report Delivery

Most BI technologies today can deliver only a single style of BI, and many customers are increasingly dissatisfied with this limitation, because it perpetuates a multiplicity of BI technologies within the enterprise that increases costs of ownership, decreases user acceptance, and limits the richness of each BI application. 

Enterprises currently need a BI technology capable of supporting any or all of the 5 Styles in any combination for any given BI application.  The only appropriate architecture for delivering all five Styles of BI is one where each style can be:

•  Mixed and matched seamlessly for users and where the addition of each new BI style adds functionality to the user’s existing reports

• Expressed through a single unified user interface to maximize ease of use and user acceptance

• Delivered on top of a single integrated backplane that unifies the metadata, security, and user profiles, thus ensuring a single version of the truth throughout the enterprise and minimizing administration and maintenance efforts by IT.

 As suggested above, MicroStrategy does just this: It delivers each style of BI equally well by offering an architecture that can deliver all 5 Styles on a single unified backplane and through a single unified user interface.  Unfortunately, other BI technologies are fragmented along the styles, thus imposing different user interfaces, different metadata and different security for each BI style. Today, most other major BI vendors are rewriting code to duplicate this single BI platform approach. 

How Information is Organized and Used

Enterprises face a challenge when striking a balance between exposing rich functionality to power users while simultaneously giving novice users a simple Enterprise Reporting environment that will not overwhelm them.

Many BI vendors solve this problem elegantly with user profiles. User profiles automatically adjust the Web interface to accommodate users with different skill levels. User profiles determine exactly what functionality will be offered to each user or user group. So, it’s relatively easy to give report designers a user profile with maximum functionality, while the report consumers receive a user profile with just enough functionality to perform their jobs easily.  Again, most BI vendors’ products can support any number of finely-tuned user profiles to accommodate the huge diversity of real people in real organizations. What’s even more important is that as user skill levels increase, an IT administrator can easily turn on more power for any given user in a centralized fashion with the simple check of a box.

Figure 1: Typical User Profiles in a BI application

How the Application Portfolio is Effectively Supported

In staffing an enterprise-class business intelligence project, there are a number of roles that need to be filled with consistent, trained, experienced resources.  For the scope of work required to bring a large organization’s infrastructure in line with the requirements of its end users, the following staffing roles need to be filled:

•  Process Manager– The process manager coordinates the interactions of all of the constituencies within the project.  The process manager works in a manner similar to a project manager, but with a deeper technical appreciation of the different systems that interact in a business intelligence application.

•   Application Architect– The application architect leads the development of the schema map that masks the complexity of the data model from the end users.  The application architect is typically heavily involved as the BI project is built and works iteratively with the Report Developers as the end-user reports are built.

• Report Developer– The report developer manages the gathering of the end-user reporting requirements and develops reporting infrastructure to support them.  The report developer typically also builds some of the more complex reports.

•  Database Administrator– The database administrator (DBA) supports the database activities required by the business intelligence application.  Typically, this involves the creation of new tables, views and indexes as needed to facilitate reporting.

• Application Developer– The application developer performs customizations necessary at the web and application layers of the BI infrastructure.  Typically, a modest amount of customization is necessary to create the appropriate look and feel for the end-users.

•  BI Alerts and Distribution Administrator / Developer  – The alerts and distribution administrator / developer creates and maintains these services.  These services provide static reporting infrastructure for disparate data sources within the business intelligence environment.

• Process Administrator– The process administrator manages the flow of BI objects across the environments discussed above.  One process administrator usually supports multiple BI projects.

• End-User Support– End-user support staff support the end-user’s demands.  This typically includes some training and mentoring of end users.

How the Alignment with Business Objectives is Achieved

Figure 2: Industrial Strength Business Intelligence Closed Loop Cycle

As the illustration above depicts, there is a need to establish a continuous loop among Monitoring, Reporting and Analyzing business results to establish and maintain alignment with business objectives. A world-class BI deployment will provide solutions to support both line and staff personnel. No one gets excluded, because they will have a choice as to whether they want to ‘subscribe’ to receive intelligence on a regular, periodic basis or be informed only when some unusual event or result that impacts them is identified.  

Necessary Technology Standards to Meet Business Objectives

The best solutions for enterprise-class business intelligence applications incorporate the following four fundamental design tenets:

1.      Pure Object Orientation: The underlying architecture was designed completely using OO techniques of inheritance and encapsulation—from its metadata architecture to the use of C++ and Java as the programming languages. The pure OO design ensures that the platform will continue to evolve and grow seamlessly with new functionality and features.

2.      Pure Web: The architecture is based on a multi-tiered structure aligned specifically with the most advanced Web design structures; including Web servers, pure HTML interfaces and application servers to ensure the application can scale to tens of thousands of users both inside and outside of any firewall with airtight security.

3.      Pure Industry Standards: The architecture exploits industry standards throughout the platform, and includes XML/XSL for report rendering and APIs; ANSI SQL for database access; HTML/DHTML, SSL and SMTP over the Internet; MDX and ODBO support for Excel access and COM for Microsoft Windows integration.

4.      Pure Open APIs: The true platform demonstrates how easily the technology can be integrated with other systems or can be customized to fit a particular need. For example, MicroStrategy engineers have exposed over 3,000 specific functions within the 7i platform for external access and customization—including all analytical functions, all system management functions and all security functions. Additionally, the MicroStrategy 7i Platform integrates with other systems using XML/XSL, Java Beans, Java Messaging and COM, and is compliant with the latest Web Services standards.

 Conclusion

 This article has only grazed the surface of issues confronting the business organization seeking to enhance the role of BI among its employees and extended stakeholders. We hope some of the points raised here are of benefit in helping readers better understand the challenges and opportunities they face when tackling BI projects. Of course, there are many excellent sources of information and guidance available in the marketplace as well as within many of your own organizations. Take advantage of any and all. We wish you much success!