During the last five years, there’s been a steady growth in the use of business intelligence (BI) software by large companies to analyze data trapped in their relational databases. BI systems, also known as data warehouses or data marts, allow these companies to detect data relationships and analyze market trends in a way that wasn’t possible with simple reporting from core database files.
The use of the tools has been primarily restricted to larger companies because the platform required to use them effectively isn’t affordable for small to medium-size companies. Here is an overview of how consultants who provide database services and who are trying to find new uses for their infrastructure can position themselves to take advantage of this need.
The real cost of BI
Executives and business managers who deal with large volumes of data agree on the value of BI to help them manage their businesses, but they find it difficult to justify the startup costs associated with the first big data mart. The cost components include database licenses, Online Analytical Processing (OLAP) software licenses, large amounts of storage, and the distribution of the results to remote locations.
Here’s where the consultant can provide significant value. If you’re providing database services, it’s likely you already have the database licenses. Moreover, mainstream database vendors (like Microsoft) include the cost of the OLAP license as part of their standard database license, making a separate purchase unnecessary.
Such consultants generally have unused storage capacity in Storage Area Networks located in their existing data centers and can lease more from their data center providers for short periods of time to accommodate data loading or cleansing phases common to all BI implementations. Finally, they would also likely have the infrastructure to distribute BI information to multi-location companies or can provide an easy means for a company to share its analysis data with customers or vendors.
Understanding the process
To provide BI services to your customers, you need to understand how end users consume BI data. Companies take existing financial and operational data stored in relational databases or other internal systems and move all of the data into a holding area where it’s cleaned—inaccurate data or missing data is completed or removed and all relationships between records in different systems is verified. This is typically done by moving all of the company data into a single set of database tables.
Then, the data is sent through a transformation process in which the necessary relationships are created to allow it to be viewed by an OLAP client. This process is called “building the cube.”
Once the cube is built, the customer will build or buy data analysis and reporting tools that can read the cube. These clients include everything from custom HTML pages to Microsoft Excel templates and macros to standard tools like ProClarity and Crystal Decisions.
Getting started as an ASP (analysis services provider)
To be able to offer BI services to your clients, you need to hire or contract with individuals who have specific BI knowledge and experience using the database, operating system, and development tools that you use in your data center.
Second, you should focus on a vertical market for your first few BI clients. Knowing the facts and measures used by specific industries is half the battle when developing cubes for analysis.
Finally, you should standardize on the back end but be as open as possible on the front end. For example, SQL Server, DB2, and Oracle all have some level of business intelligence capability built into their core product. SQL Server is the richest out of the box, but Oracle has more third parties that can hook into it on the back end.
On the front end you need an easy, inexpensive, per-company access capability that doesn’t require per-user licensing. (You’ll probably have to develop this yourself or buy a site license for a lesser known analysis and reporting tool.) If you can rely on your customers to have a recent version of Microsoft Excel (2000 or higher), you can use it as a rich client by providing macros and templates that expose the data in the cube. Once you’ve developed the capability to deliver hosted BI solutions, you’ll find lots of customers waiting to take advantage of them.
What’s the biggest obstacle to BI?
Are consultants taking advantage of this trend toward providing BI services for their clients? What’s the best approach? Post your comments below.