More and more organizations that are doing basic reporting and data analysis are seeking to expand the business intelligence process to get more of the benefits of BI and advanced analytics.
Making that happen isn’t so easy, though. In interviews at this month’s TDWI World Conference in Orlando, Fla., a half-dozen attendees detailed ongoing efforts to broaden the use of their data warehouses and BI tools. But they cited a variety of familiar challenges, including data quality problems, internal IT limitations and the need to sell business users comfortable with spreadsheets or other standalone tools on the potential value of using an enterprise data warehouse and BI system.
In many cases, overcoming those challenges can take years. That’s the situation facing Ranjit Pothuru, associate manager of data strategy and data warehousing at Everest Re Group Ltd., an insurance holding company with its main operations office in Liberty Corner, N.J.
Everest Re, which offers reinsurance services as well as property and casualty insurance policies, built a data warehouse six years ago. But thus far, the warehouse has been used primarily as a data source for regulatory compliance reporting. And before the company can move forward on a plan to turn it into more of a BI and analytics engine, there are “a huge number of data quality issues that we have to tackle” across the organization, Pothuru said.
The data quality fixes will include improving data cleansing, establishing data governance rules and potentially launching a master data management program – all of which could take up to two years to fully put in place, according to Pothuru. In addition, he said that Everest Re’s IT team will have to spend time reaching out to end users who have done reporting directly off of the company’s mainframe systems for decades and “still aren’t entirely sold on the reliability of the BI system.”
Pothuru, who was hired last year as Everest Re’s first dedicated BI and data warehousing worker, said compliance reporting has to remain a key focus area. “But on top of that, we don’t want to be left behind on analytics,” he said.
The expanded BI implementation will revolve around financial data at first, but Pothuru said plans also call for providing the company’s actuarial workers with BI tools that will enable them to analyze how different insurance products are doing in the market.
Taking the business intelligence process out of ‘the bowels of IT’
A similar push to get more out of the BI process is on at Minneapolis-based Fairview Health Services, a nonprofit organization that operates seven hospitals and dozens of other medical facilities in Minnesota.
Like Everest Re, Fairview has a data warehouse that for years “has been hidden away in the bowels of IT” and used primarily for financial reporting, said Jack Weber, IT manager on the healthcare provider’s BI and data warehousing team. But now, “the mentality of the business has changed,” he added. “There’s a greater realization that the data is actually useful to us.”
Fairview is looking to expand the use of the data warehouse to areas such as BI for patient care, in an effort to become more proactive on treatments and ensure that its medical practices meet or exceed industry standards. To help speed up the delivery of new analytical capabilities, the company is moving to implement agile BI techniques, which Weber said is driving a need “to really change how my team thinks” about BI development and working with the business side.
Jack WeberIT manager, Fairview Health Services
Scripps Networks LLC, which owns HGTV, the Food Network and other cable TV channels, built an Oracle-based data warehouse three years ago, initially to report on advertising sales data. For the past 18 months, the company has been working to develop a broader BI strategy – an initiative that finally is resonating with business users, according to Douglas Wielfaert, a data analyst in its enterprise BI group.
“Suddenly, after three years [total], the light has turned on,” Wielfaert said.
Scripps Networks is now extending the BI process to functional areas such as program planning, where workers previously relied heavily on spreadsheets. The new worry, Wielfaert said, is that the BI floodgates will open as users request new tools and capabilities – potentially stretching the BI group’s ability to keep up.
Philip Russom, an analyst at The Data Warehousing Institute, which organized the conference, said that many companies have mature data warehousing and reporting processes but have yet to move up to more advanced BI and data analytics systems. But interest in doing so is on the rise, Russom said, noting that the number of people enrolling in TDWI’s analytics training courses “is way up.”
From spreadsheets to a more structured business intelligence process
American Access Casualty Co. made the leap over the past 18 months. The Oakbrook Terrace, Ill.-based auto insurer didn’t lack for data before, said Kevin Rooney, formerly its CIO and now chief strategy officer. But the company’s policy pricing analysts “were buried in spreadsheets,” and the data they were using was siloed and not always timely, Rooney said during a keynote speech.
To try to create a more structured BI process, American Access deployed a cloud-based pricing analysis system built around Kognitio Ltd.’s analytic database and software from Quomation Insurance Services Inc. The BI system combines internal data and U.S. census information with hundreds of millions of external auto insurance pricing records, and Rooney said a 60-day proof-of-concept project pointed to an immediate profit boost of $175,000 – a big number for American Access.
But the insurer is just now reaching the point of filing proposed new pricing rates in Nevada, the first state where findings from the BI system are being applied. Rooney acknowledged that American Access hit some bumps in the road – for example, the pricing analysts had trouble dealing with all of the data suddenly at their fingertips. “Getting the wrong information faster wasn’t exactly the intention,” he said.
To get things back on track, Rooney’s team worked closely with the analysts and got more business executives involved in the BI project.
“The good thing is we had established a strategy up front, and everyone was committed to that vision,” he said, adding that doing so “gave us the confidence that this project would get done” despite the hurdles.