Athena IT Solutions founder Rick Sherman makes this point in a SearchBusinessAnalytics.com piece and offers some great suggestions for maximizing BI performance on a tight budget. Among his tips:
- Employ a hub-and-spoke architecture. In this architecture, a data warehouse serves as the hub, for loading, cleansing and storing large volumes of information from operational systems and other data sources, while data marts are used for specific BI uses such as processing data to populate dashboards and reports and answering online analytical processing (OLAP) queries. To make it even more scalable, Sherman says you can create smaller data marts or cubes within data marts.
- Invest in a database administrator instead of asking an application programmer to do DBA duty. They'll be more likely to create databases that can support the required levels of BI performance.
- If you want to deploy a variety of BI tools to satisfy different business needs (and this is a good idea), you'll often spend less if you get as many of the tools as possible from the same vendor. Another option (albeit one that Sherman doesn't mention) is using software-as-a-service for some of these supplementary apps.
- Establish development standards and create templates or style sheets for developers. In addition to freeing up developer time, Sherman says standards help ensure more consistent apps.
- Spread out the databases as well as your extract, transform and load (ETL) processes and BI querying across various physical or virtual servers, taking into consideration how they fluctuate throughout the day, week and month.
- Get as much memory as you can afford, and leverage in-memory analytics and ETL caching technologies when possible. In-memory analytics work best with a 64-bit architecture.
Consolidating BI tools when and where possible is another good money-saving suggestion. Duplicative technologies rarely make sense, and it's obviously not cost-effective if different groups of users perform similar analyses with different tools.
In a post I wrote about a year ago, I cited an article that offered the example of Allstate Insurance Co., which shrank its number of data warehouses from 13 to two and eliminated two-thirds of its BI tools by scrapping redundant ones and those no longer popular with users. The result: lower software license and support costs, and user training costs. The move also made it easier for different departments to share data.