For nine years, Ryan Fenner worked in a group that developed and managed business intelligence systems at Union Bank outside the IT department's control. Now he's helping clean up some of the consequences of such efforts, as the technical leader on a data warehousing project aimed at bringing BI and analytics data generated by that group and others like it into the bank's corporate IT fold.
The project isn't designed to fully do away with shadow IT operations at Union Bank, which is based in San Francisco and has 400-plus branch offices in seven states. Fenner, who moved into the IT department in late 2012, said some of the groups will remain in place and continue to build BI and analytics applications for business units. The goal is to get the key data they're working with into centrally managed data marts, and eventually into an enterprise data warehouse (EDW), so it can be properly cared for.
Fenner learned the need for such care firsthand several years ago, when the bank's risk and compliance department flagged his shadow IT group for data quality problems. He said the various DIY operations sprung up because there wasn't a strong central focus on BI and analytics -- but the informal processes ended up creating data consistency and accuracy issues. "A lot of innovation comes out of shadow IT, and they provide a lot of value to the business," Fenner said. "But you do need governance on that data."
He started his career on a traditional IT track, graduating from college with a degree in management information systems and initially working as a business analyst and then a systems analyst. His first job at Union Bank involved rewriting Microsoft Access desktop applications to run on a SQL Server system. Along the way, he became familiar with running queries and building reports in SQL Server -- and then a shadow-y opportunity came knocking.
Do what you need to do
Business executives in the bank's consumer lending unit were looking for better BI data to aid in decision making, and they tapped Fenner and two colleagues to start producing reports for them. He said the execs gave the new team "tons of cash" to fund its work and free rein on how to proceed: "They said, 'We need questions answered and we don't care how you do it -- just do it."
A lot of innovation comes out of shadow IT. But you do need governance on that data.
Six months later, the group was expanded to support the company's entire retail banking business. Eventually, it cast quite a long shadow, employing more than 100 people at its peak. That pales in comparison to the size of Union Bank's IT department, which currently has about 1,300 employees. But it was a sizable operation.
And it was fun to work for, according to Fenner. Using SQL Server as their BI platform, he and his co-workers got to do "cool things" in direct support of business needs, with a heavy focus on analyzing sales and marketing data. But after a management change, the work "morphed into more of a traditional IT role," he said. Meanwhile, the IT department also got new management and became more focused on supporting BI processes. So Fenner decided to leave the shadows behind and take a new job in IT as a vice president and enterprise data solutions architect.
First priority: Getting a handle on analytics data
Officially, he's in charge of enterprise architecture. But, for now, architecture strategy is mostly taking a back seat to the tactical concerns of the data warehousing project. The project team deployed an initial data mart for Union Bank's consumer loan servicing department in late April, building it on SQL Server using data warehouse development tools from WhereScape. Fenner said he expects the reporting system to be migrated to Pivotal Software Inc.'s Greenplum analytical database later this year, again with WhereScape's tools. Other data marts will follow, and the bank ultimately plans to construct a full-fledged EDW on Greenplum that will incorporate the marts and several Oracle-based data warehouses built by the IT department.
More on managing BI and analytics data
Watch a video Q&A with IT exec Sam Strum about managing data warehouse and BI projects
Get tips from consultant William McKnight on how to avoid data warehouse performance problems
See why consultant Wayne Eckerson says big data vendors should stop bashing data warehouses
Fenner is applying some familiar concepts from his shadow IT days to the initiative. For example, he isn't a big fan of traditional requirements gathering and waterfall development. In his old group, he said, "we went out and sat with the business people and just built the applications," using Agile development processes. The data warehouse project team is adopting a similar approach.
Things aren't all sweetness and light on the project. The bank isn't looking to pull all of the analytics data generated by shadow IT operations into the new data marts and the EDW -- only information that it views as an enterprise asset. Fenner gets to make the call on what qualifies, even if current data owners would prefer to keep it to themselves. "There are some groups that don't want to work with us," he said. "We have to make sure we prevent that kind of thing."
But if everything works as planned, Fenner added, the end result will be stronger data management that improves data quality without shackling the analytics capabilities of BI teams embedded in business units -- still enabling them to deliver "beautiful innovation and a lot of business value."
- Successfully Managing Data in a Cloud Data Warehouse –Informatica