What's the best way to prevent the finger pointing and animosity that plague many data warehouse projects? The best way, if not the simplest, is to rotate data warehouse team members through various jobs so they can experience the project from several viewpoints, according to data warehouse gurus who spoke this week at The Data Warehousing Institute (TDWI)'s 2003 World Conference in Boston.
Of course, that's a suggestion that sounds great on paper but can be tough to implement, as Glenmoore, Pa.-based Kathleen Garrity, a business systems analyst for the Vanguard Group Inc., pointed out. Like many attendees, she is grappling with the personalities and people that affect her data warehouse project the most.
Garrity is currently preparing a new team to implement and maintain a Cognos data warehouse, running on an IBM DB2 database. One problem Garrity is facing is that new mainframe employees are being forced to learn object-oriented programming. There's no easy way to train them, Garrity said.
"We're trying to do a rotation," she said. The problem, Garrity said, is that "we have two new people without a clue, and a lot of people who know a lot but can't spend time training." Sometimes, the experienced programmers don't want to be mentors, she added. "It's like a snowball going down a mountain. It takes on a life of its own," said Garrity, echoing comments from many of her colleagues, whose data warehouse projects have sped out of control.
The method of rotating employees is a proven one, though, and has been a success at companies such as Proctor & Gamble, the international manufacturer of personal care products.
Conference attendee Andrew Towle, project manager of IT sales and marketing for Quest Diagnostics Inc., is heading a team that is building a PeopleSoft Inc. data warehouse, using an Oracle database. Towle said he would like to see each person on his data warehouse team experience the project from the viewpoint of a database administrator. "We'd really end up with a decent team," said Towle, who is based in Norristown, Pa.
A data warehouse plan for the ages
Maureen Clarry and Kelly Gilmore, co-founders of Connect: The Knowledge Network, conducted a full-day TDWI conference workshop on leading and organizing a data warehousing team. They offered a framework for addressing conflict, finger pointing and communication issues within an organization.
They referenced the Denison Model, a tool based on 15 years of research conducted by Daniel Denison at the University of Michigan's business school. Intended to help organizations understand and design data warehousing projects, as well as improve leadership practices, the model offers four categories that can help define the success of a data warehouse project.
Mission: "Do we know where we are going?" A stable project element that accounts for the direction, purpose and blueprint for your team or organization.
Adaptability: "Are we listening to our customer and the industry?" A flexible trait that is affected by industry patterns and trends.
Involvement: "Do we have buy-in?" A flexible trait that includes employee commitment, ownership and responsibility of tasks.
Consistency: "Can we execute?" A stable trait that applies to required systems, structures and processes for team success.
"We have to be flexible, but we have to be stable," Clarry said. "It's a balancing act." Ultimately, people have to understand one another and what's required of them.
For More Information: