Earlier this year John Kitchen, SVP and chief marketing officer for Datawatch Corp., wrote a guest opinion for IT Business Edge in which he discussed his company's approach to dealing with what he called the toughest business intelligence challenges: BI is too expensive, it doesn't tell organizations what they want to know and organizations don't trust the data. He advocates mining data from reports found in existing enterprise systems. He explained:
Reports can be mined to eliminate the need for a BI system in relatively simple applications or to supplement a traditional BI system by providing users with a tool to perform analysis that is not supported by the traditional system. Data can also be mined from reports and directly uploaded to a data mart or data warehouse, again reducing the need for new programming or new, complex data queries.
That sounds like the approach used by Ohio's Miami University in a project to determine the cost of its summer courses and to boost summer enrollment. The school's ERP system already held much of the information, but not in a format conducive to data analysis, reports Campus Technology. Instead of buying new software, the school opted to work with software it already had to create a data mart and build cubes to query the data mart.
Consultants were brought in to help, and the project was completed in three months and with a budget of less than six figures. This was even more impressive considering the university had only one BI specialist, a data modeler, on its staff. I thought it was great to see an organization that put together an effective solution with existing technology rather than simply sighing and writing a big, fat check. Sometimes you don't have to reinvent the wheel, you just have to figure out a new way to roll with it.
Assistant Director of Business Intelligence Phyllis Wykoff did a number of things right in the implementation. None of these steps are exclusive to data marts. Rather, they are a good idea for any BI project. Among them: