At the Business
Intelligence RoadShow held in Toronto on Tuesday, the university’s senior business analyst and project manager, Debbie Weisensee, discussed the challenges she faced when getting executive buy-in for BI, such as overcoming sticker shock and the initial resistance to a long-term IT project.
It took six months to persuade upper management to invest time and money in BI at McMaster University. In the end, Weisensee, said she presented a business case around what it would cost them if they didn’t invest in BI.
McMaster is a mid-size university in Hamilton, Ont., with 19,984 full-time students and 1,189 academic staff. But efficiency has not been its strong point. For example, 60,000 account statements are printed out every month, which are then re-keyed into spreadsheets.
Part of this inefficiency is due to the fact it has several disparate systems with data stored on segregated platforms. This presents problems with data extraction and integration, as well as reporting. If someone needs to pull data out of the system from five years ago, for example, they would have to go to the IT department and put in a special request.
“If we can’t get it when we need it, it’s not valuable,” said Weisensee.
The university also uses different tools for data extraction, which has resulted in a reliance on specialized knowledge. In terms of reporting, excessive time is spent defining, assembling and verifying data – and not enough spent on planning, analysis and innovation. Decisions are being made on incomplete data.
“Most of our decisions are gut-based, not evidence-based, which is where we want to go,” she said.
The solution was to develop an enterprise-wide data warehouse that would be deployed incrementally along with Web-based query and reporting tools using SAS’s Intelligence Value Chain.
But most data warehouses don’t have a high success rate, Weisensee said. So she identified 71 common information needs across the university as part of the business case for BI.
The anticipated outcomes of this rollout include a single, standardized source of data, a reduction of time spent re-keying data, and the ability to share information and customize reports. Two marts and 30 reports have been built to date (but are still in a pilot phase).
BI will be rolled out mart by mart, and Weisensee expects to roll out something every quarter for the next three to five years – as long as she can prove it’s adding value.
The next phase, which will start in September, involves developing balanced scorecards and performance indicators. Weisensee wants to expand beyond financial data, to focus on areas like enrolment and student information. Predictive analysis would allow the university to make predictions with its data; for example, student interest in a particular course could be matched up with the number of professors teaching that course.
Two distinct customer requirements for BI are emerging, said Joel Martin, vice-president of enterprise software and business alignment solutions with IDC Canada. Core business analytics involves tracking, analyzing and delivering information, while predictive business analytics involves hypothesizing, modeling and acting.
Better intelligence is needed to capitalize on investments, he said, to further automate business processes or build in “triggers” that will help organizations react more quickly.
The top business problems taking a priority this year are the need to integrate with business partners, introduce products or services faster, offer more customer service solutions, service existing markets and increase productivity. And the No. 1 investment in application extensions, he said, is business intelligence and data warehousing.
Most large companies have different types of applications from different suppliers, and ongoing application interoperability will be the catalyst for BI this year and next, he said. Organizations need to look beyond simply mining databases and figure out how they can make the most out of their investment. “It’s not just another technology upgrade.”