Cinco tips para una exitosa implementación de Inteligencia de Negocios (Inglés)

5 Tips to a Successful Business Intelligence (BI) Application

• Tip #1: Clearly Define the Project Plan

A good project plan outlines the criteria for success and establishes a road map for achieving that success. In addition, a clearly defined plan identifies project milestones and allows clear measurement of progress throughout the project life cycle.

Achieving success is often a matter of meeting (and exceeding) expectations of the user community and the project sponsors. It is important to set realistic expectations with regard to application functionality, query response time, delivery schedules, and  costs of implementation. The project team should strive to maintain a project scope consistent with these expectations.

Delivering a successful business intelligence application requires a careful balance of three important factors: user requirements, query response time, and maintenance overhead. Clearly there are tradeoffs. For example, one could create aggregate tables to improve query response time, but only at the expense of additional database maintenance.

There will be many design decisions throughout the project. Each should be made with an understanding of how these three factors are affected. The project team should have a clear idea of the relative priority among the three factors.

• Tip #2: Design an Appropriate Data Model

Business intelligence applications differ significantly from transactional database applications.  Unlike transaction processing applications that take advantage of well-understood business processes, business intelligence systems add value by allowing innovative ad hoc analysis of information.  It is essential that the project team build an  application on top of a database designed to accommodate this valuable type of analysis.

The most successful business intelligence projects emphasize the design of an appropriate, flexible, and maintainable data model.  The business intelligence system, and the underlying data model, must be intuitive for end users.  Users should be able to analyze the data using known business concepts and terms.  The database must also be designed with performance in mind.  Because the patterns of data access can vary with business intelligence applications, the database should be designed to optimize query performance in an ad hoc environment.

• Tip #3: Ensure Quality Through Comprehensive Testing

Resources for data extraction and transformation are often underestimated in business intelligence projects.  The complexity of the transformation effort increases exponentially with the number of data sources to be integrated. Successful projects are planned with the challenges of data transformation in mind.  To be successful, the transformation routines must be production ready and accomplish all necessary data extraction and transformation tasks. The batch processes must complete all processing in the allotted batch window and require no manual intervention once in production.

In order to be production-ready, the transformation routines along with all other system components should be thoroughly tested.  A comprehensive testing phase ensures production "readiness" in the application.

The most successful business intelligence systems include at least three testing phases: unit testing, system testing, and user acceptance testing.  The degree of formality is less important than ensuring that the following topics are addressed: user feedback on functionality and usability,suitability of the data model to query requirements, validity of reports and objects, accuracy of data and results, user concurrency, RDBMS query throughput, and timely completion of transformation routines.

• Tip #4: Create a Positive End-User Experience with Training and Phased Rollout

A two-part strategy for end-user training helps improve the overall user experience with any new business intelligence application.  First, end users should be trained on the data they will use.  Often, a new business intelligence application introduces data and metrics that are unfamiliar to users.  In these situations, users require some familiarity with the data before they even touch a mouse.  Second, end users should be trained on the application they will actually use.  Training is most effective when users can relate to the business terms and scenarios that are meaningful in their work.

The most successful business intelligence projects feature a phased rollout strategy.

A controlled rollout typically consists of initially rolling out to a subset
of users.

This introductory period provides an opportunity to test the application,gather  feedback, test concurrency, tune the hardware and software, and add or remove functionality.  The introductory phase is then followed by the production phase.

For some projects, it may be necessary to introduce new users one group at a time. This incremental rollout allows the project team to refine the training process and monitor the production system as usage increases.

• Tip #5: Ensure Success with A Balanced, Well-trained Project Team

The business intelligence project team should possess strong skills in all of the business intelligence tools to be deployed.  Application developers, system architects, and system administrators should have the skills to develop metrics, manage metadata, and deploy web-based query, reporting, and analysis applications.

The DBA staff is responsible for a number of critical tasks in the business intelligence project.  Skilled DBAs have practical experience in managing production databases. They have experience in RDBMS-specific configuration, backup and recovery strategies, performance tuning, system monitoring, data loading and storage strategies.  In query, reporting and analysis applications, DBAs should also have an understanding of multidimensional design concepts, query access patterns, indexing strategies, partitioning techniques, and pre-aggregation strategies.

For some business intelligence projects, you may consider the use of
specialist consultants.  The use of consultants who are experts in the RDBMS platform being used for the project is extremely important.  Much of the technical success of a business intelligence project depends on the ability to maximize the performance and capacity of the RDBMS.  RDBMS consultants are available through your RDBMS vendor or various third-party channels.

We hoped you enjoyed "Five Tips to a Successful BI Application."

If you would like to learn more about MicroStrategy and business intelligence:
Take the quick tour "What is BI?" http://www.microstrategy.com/
Sign up for one of our daily interactive online seminars: http://www.microstrategy.com/Events/Online_Seminars/
Order a free eval CD of MicroStrategy http://www.microstrategy.com/FreeEvalCD.asp

Laura Bodwell
Manager, Direct Marketing
+1.703.744.6318 (ofc)
+1.703.599.4792 (cell)
Esta dirección de correo electrónico está siendo protegida contra los robots de spam. Necesita tener JavaScript habilitado para poder verlo.
1861 International Drive
McLean, Virginia 22102
MicroStrategy - Best in Business Intelligence TM

Te gustó? Compártelo !

Tweet

Whatsapp