When Florida State University (FSU) undertook a large-scale ERP implementation back in 2004, the focus was on improving the school’s financial and human resources transactional applications.
Business intelligence (BI) reporting and data analytics were not top of mind, according to Michael Barrett, chief information officer at the Tallahassee, Fla.-based university.
But times have changed, and in 2007 Barrett decided it was time to standardize all of the school’s reporting, dashboarding and analysis capabilities on one BI platform.
Until then, most university employees relied on the reporting capabilities prebaked into their PeopleSoft ERP suite. The school also used Business Objects to perform ad hoc analysis of enrollment and admissions data that was collected and stored in a separate database.
But the weekly and monthly financial and human resources reports proved insufficient, Barrett said. Workers wanted access to daily (and sometimes near-real-time) reports to better track spending against budgets, and tools to perform tasks like headcount analysis.
And the Business Objects deployment, while functionally adequate, was not integrated with data from the school’s PeopleSoft suite. “That was really quite separate from the main ERP system,” Barrett said.
Florida State’s IT department was essentially spending time and manpower supporting two systems -- PeopleSoft and Business Objects -- neither of which was meeting the school’s BI and data analytics needs.
Time was right to standardize BI functions
After evaluating a number of options, Barrett decided to standardize on Oracle’s Business Intelligence Suite Enterprise Edition (OBIEE) Plus. A number of factors played important roles in the decision, he said, including the fact that Oracle acquired PeopleSoft in 2005.
Since the acquisition, Oracle has built connectors to integrate many of its BI functions into the PeopleSoft ERP suite, meaning less work for Barrett’s team than would have been the case if they had opted for another BI vendor.
More importantly, though, OBIEE “met 80% of our needs right out of the box,” Barrett said. Just the remaining 20%, including extending real-time analytics to the school’s payroll processes, for example, would need to be custom built.
The Oracle BI tools were also easy to use and could integrate with Microsoft Excel, he said, meaning that he would not have to spend inordinate amounts of time retraining the 1,500+ Florida State workers who would regularly use the system.
Despite all the benefits, however, the deployment itself would consume significant IT resources and last several years. It is still ongoing.
Multi-phase BI software rollout
Rather than try to deploy all of the newfound BI and analytic capabilities at once, Barrett decided that a multi-phased deployment was the best approach. Starting in 2007, the first phase focused on 13 core subject areas, including human resources and financials, he said. By February 2008, he and his team had created and deployed 26 dashboards using the new Oracle BI suite.
Since then, Barrett said, OBIEE’s self-service capabilities have also enabled him to shift the burden of report creation and analysis from IT to end users. Subsequent phases have focused on extending analytics to Florida State’s grant management work.
Next, Barrett plans to roll out OBIEE to the school’s admissions workers, who previously used the siloed Business Objects deployment for ad hoc analysis of enrollment data.
Barrett said the multi-phased approach allowed his staff to deploy the Oracle BI suite in manageable chunks, and he recommends the approach to others considering standardizing their BI infrastructure on one vendor’s technology. In addition to reducing the stress level on IT, it also provides a steadier stream of so-called quick wins, which helps keep business-side sponsors engaged.
“You’ve got to be reasonable about what you’re going to deliver and try to deliver it in reasonable chunks,” Barrett said. “You’ve got to manage it like any project so it doesn’t drag on.”
He also said it is important to understand the amount of work involved before getting started and to set expectations accordingly. There are, for example, hardware and infrastructure implications to consider when rolling out an enterprise-wide BI system. DBAs will need to spend time doing SQL analysis and tuning to fix bottlenecks in the system. And, as easy as some BI tools are to use, end users will need to be trained and supported as problems arise.
Currently, nearly 1,600 Florida State employees from “literally every department in the university” use the Oracle BI suite, Barrett said. “People have a lot more data at their disposal, and it’s a lot more timely.”
In addition to improved reporting and analytic capabilities, the university has also saved around $350,000 in maintenance/license fees and support costs as a result of standardizing its BI technology, according to Barrett.
“There is definitely a cost,” he said. “But it was well worth it.”