Successfully implementing a business intelligence (BI) solution is a challenging, complex and valuable endeavor. Overcoming challenges and complexities requires a combination of the right implementation strategy and technology. Understanding the interdependency of these factors is critical; choose the wrong strategy for the selected technology or the wrong technology for the selected strategy, and your chances of success are significantly decreased. Find the right combination and you can give your organization the ability to make better, more confident decisions and enjoy the competitive edge this ability brings.
There are many successful combinations of strategy and technology, but an iterative, organic approach using Microsoft technologies is becoming increasingly popular. Microsoft's tools are improving, and more BI professionals are adopting Microsoft technologies as a viable platform for an iterative deployment approach. To appreciate why this strategy and technology combination is gaining popularity, it is important to understand how an iterative, Microsoft-based solution overcomes the key challenges inherent in all BI solutions.
Organic Business Intelligence
BI is truly a journey, not a destination. By its very nature, BI provides value only if it provides meaningful information and insight into an organization's current business metrics and key indicators, which frequently change. A successful BI solution must be nimble in order to accommodate these changes, or it will quickly become obsolete.
Long development cycles are deadly in this changing environment. Not only must the physical solution be nimble, the deployment plan must consist of relatively short-cycle, discrete deliverables that may be reprioritized as needed. Making the business wait for a previously considered "nice to have" future capability that has suddenly become critical is not the right answer. The greatest value of BI is its ability to deliver the information the business needs when it needs it.
A successful BI solution must also be robust and scale to meet the demands of new subject areas, users and functionality without falling out of performance parameters. If users find that their needs cannot be met by an underpowered system, they'll develop their own solutions and your BI investment will be lost.
Though it is possible to successfully implement this organic BI environment using different technologies, Microsoft's SQL Server platform, Office Suite and available third-party visualization products natively understand and support iterative deployment cycles and changing requirements. Increasingly, BI implementers are considering Microsoft technologies as a viable platform for large enterprise BI solutions.
Plan for Tomorrow but Build for Today
It is critically important to have a vision and implementation roadmap for your BI solution. These are your guides in evaluating your deployment schedule and measuring your progress. However, it is a mistake to attempt to build your entire vision in a single deployment.
Plan short, regular deployment cycles and address a reasonable set of current needs to manage both speed of delivery and cost. Since this strategy allows you to deliver value quickly and roll out incremental, additional value over time, development cost is spread over many cycles, producing a positive cost-value ratio. Typically, this is preferred to a front-loaded strategy of implementing a more comprehensive solution in a single, longer cycle.
This iterative deployment strategy also allows for the reaffirmation and adjustment of the BI roadmap as part of the deployment planning process. The result is an organic implementation process that produces a BI solution closely modeling business need.
This is not a Data Warehouse Project
A data warehouse is a component of BI, but is not the solution itself. Many BI implementations fail because the primary focus is the data warehouse, not the overall value of timely, meaningful information. Data warehouse projects place the emphasis on the warehouse, which can be a very costly mistake.
A properly formed and managed data warehouse is critical to your success, but a successful BI solution must always be evaluated by its specific business value.
Users perceive the value of BI in the visualization layer as a scorecard, report or interactive analytic capability. Though the data warehouse often provides the information necessary for these visualizations, the warehouse itself is uninteresting unless it provides credible, conformed and timely information that can be understood and used by the business.
This value test also allows us to limit the data warehouse and all the data acquisition and conditioning work to only the data required to support the business need. The warehouse, as a component of the overall solution, will evolve over time. There is no need to load or model all the organization's data today, even if we think it might be useful tomorrow.
Abstraction in Business Intelligence
Abstracting business users from the technical environment is a key component of BI. The visualization of BI via a report or analytic tool should depend upon a logical business model for the users and applications. This allows for interaction without specific knowledge of the physical architecture as well as architectural change over time with minimal or no disruption of existing reports and applications. The stability of the visualization layer allowed by this abstraction is a fundamental aspect of the success of your BI implementation.
Rules of Thumb for Successful BI
Though each BI implementation will have unique characteristics and challenges, some rules of thumb can guide you in making fundamental design and approach decisions.
Data Acquisition. Data sources will change over time, so design data acquisition processes to be modular and replaceable. As a consumer of these sources, your BI solution must pull data from new or legacy systems with as minimal impact as possible on other solution components, particularly on the data warehouse and staging areas.
Centralize your data-conditioning business rules. Regardless of the number of sources you are using to populate your data warehouse, it is best to use a single tool for data acquisition to enhance consistency and reduce the cost of maintaining business rules. If you're working with Microsoft technologies and want to minimize your cost of licensing ETL tools, SQL Server 2000 includes Data Transformation Services (DTS). SQL Server 2005 replaces this capability with the greatly enhanced Integration Services (SSIS). Third-party vendors such as Business Objects, Harte Hanks, Trillium and Informatica, among others, also provide very strong ETL capabilities for managing SQL Server data warehouses.
Data Management. Use conformed dimensions with subject area-specific data marts. This architecture allows for a centralized, managed data warehouse while supporting targeted, tailored delivery of data to discrete users and user communities.
Multidimensional cubes in Microsoft's Analysis Services (SQL Server 2000 or 2005) provide terrific query performance for ad hoc analysis as well as structured reporting, particularly where aggregations or comparative or trending analysis are required. Though cubes require a different skill set than relational stores, these pre-aggregated, business-modeled marts are extremely valuable in a mature BI solution.
No user or application may directly access physical tables. All user or application access to data should be via a logical layer. Whether in the form of relational views, multidimensional virtual cubes or the Unified Dimensional Model (SQL Server 2005 Analysis Services), this logical layer abstracts users and applications from the physical environment, allowing continual change in the technical infrastructure while maintaining stability in the business environment.
Visualization. One tool does not fit all. Business users will require different tools to meet a variety of needs. For instance, while virtually all users will have a reporting requirement, a smaller number will need ad hoc analytic capabilities.
Also, scorecards and dashboards can be very effective reports as well as launch points for further analysis. In order to increase adoption of your BI solution, deliver the information in a variety of ways.
Putting it all Together
Delivering such a capability requires an appropriate combination of implementation strategy and technology. The increasingly popular iterative, organic implementation approach provides value early in the cycle and additional, relevant value with subsequent incremental deployments while spreading cost over time. This strategy, implemented on Microsoft's SQL Server platform, is a strong combination for success.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access