LESSON - Key Requirements for Enterprise Business Intelligence
By Mark LaRow, Vice President, Products, MicroStrategy
In the past, companies deployed BI applications as departmental solutions, creating islands of BI for a distinct user population and database. Early success with departmental-level BI has led to greater demand to expand BI across the organization. As enterprises open up their ever-growing data assets to everyone in the company, they are now faced with the challenge of doing so in a cost-effective and scalable way.
The solution to this challenge is enterprise BI standardization, where one BI platform delivers a single version of the truth through a unified user interface to all people across the company. There are five key requirements for enterprise BI:
1. Support for Five Styles of BI
As a company moves into enterprise BI, it must have a technology that can seamlessly deliver all five styles of BI with a single unified architecture and metadata. The five styles of BI are:
- Scorecards and dashboards
- OLAP analysis
- Advanced and predictive analysis
- Alerts and notification
Equally important is the need for a single technology to provide all five styles as plug-and-play components on a single unified backplane and through a unified user interface. The unified backplane ensures that all styles of BI share common services and automatically build upon one another. The unified user interface ensures a common experience for all users, allowing them to seamlessly move between styles of BI without cumbersome application switching.
2. High User Scalability
An industrial-strength BI architecture must support thousands of users, interactively accessing the same reports and scorecards. While many BI vendors cite large user populations, they do not mention that these high numbers are aggregations of many small, isolated BI applications, or that they require massive server farms.
To achieve the highest user scale, companies need a BI technology capable of reaching all business users through the interface of their choice; a zero-footprint Web interface that is instantly deployable to geographically distributed users; and a dynamic caching architecture. A dynamic, multi-level caching architecture provides increased performance as more people use the system, making each user’s experience faster and more interactive, while imposing the absolute minimum load on database resources.
3. High Data Scalability
Companies are faced with an explosion in the volume and scope of data they collect and generate. The myriad of systems, including ERP, CRM, Web sites, sales force automation, SEC compliance, and supply chain management, contributes to this massive data influx. Enterprise BI must be able to access all these databases to be truly effective.
A ROLAP (relational OLAP) architecture is necessary to access all data to the full depth and breadth of the largest databases. While some reporting-specific BI technologies can access very large databases, they sacrifice user interactivity. Reporting tools can access large databases, but only with relatively simple queries and without any real user interactivity. On the other hand, traditional OLAP tools can provide a high degree of user interactivity and exploration, but only against comparatively small amounts of data, stored in cubes, that is extracted from the relational databases. Relational OLAP architecture provides access to the largest databases with the highest interactive performance.
4. User Self-Service
As the population of users and amount of data in the enterprise grows, user self-service becomes absolutely critical. Enterprise BI requires levels of self-service that go beyond ad hoc report design. To achieve broad-based self-service, companies need a BI technology that offers “what you see is what you get” (WYSIWYG) design, eliminating user training in report design. Even more effective for user self-service is a BI technology that can drill anywhere, allowing users to seamlessly drill through any combination of data in the data warehouse.
5. Automated Maintainability
Finally, cost-effective enterprise BI requires the ability to update all reports automatically, reflecting continual changes to underlying business definitions, business structure, and database structures. With a small BI application, the effort required to maintain reports manually is modest. However, with enterprise BI, report maintenance becomes the dominant issue in total cost of ownership, and automated instantaneous updating becomes a critical requirement.
Only BI technologies with a dynamic metadata architecture can ensure that all reports are updated instantly and automatically, with a single version of the truth, at any scale. Reports are immediately brought up to date whenever any underlying object changes, requiring fewer IT personnel to deliver the maximum amount of BI.
Enterprise BI has the power to deliver timely business information to everyone in an organization. Given the changing landscape of reporting, analysis, and monitoring requirements, traditional reporting tools and non-integrated BI products are reaching their limitations. As companies consider the benefits of enterprise BI, they must carefully choose a technology that can meet the most demanding requirements and enable thousands of business users to simultaneously access terabytes of data with mission-critical reliability.
This article originally appeared in the issue of .