June 13, 2002

The Dream Team

Assembling a team with the right mix of hard and soft skills for your business intelligence project is critical to its success

By Patti Bunker

You've probably read enough articles and worked on enough projects to know that projects fail for a variety of reasons. You even know that a bad project team can be a major contributing factor to project failure. A lot of factors go into a good project team: Does the team gel? Do members share a common work ethic and goal? Are they able to perform the tasks at hand?

A good team is even more important for the success of business intelligence (BI) projects. Good BI projects, by their nature, are designed to meet strategic business goals and, therefore, require much more involvement and input from people on the business side of the organization. So how do you find the right people for your BI application development team?

Assembling a solid BI project team isn't always easy. You should be open to finding the right people from both within your IT group and from the business side. If your budget allows, you may need to look for resources outside your company (which can be an advantage in terms of getting people without agendas, loyalties, or expectations).

Although technical skills are important, if your team doesn't have the necessary soft skills to handle both the business and technical aspects of the project, it's doomed to failure. In-depth technical interviews can determine if the folks on your project team can actually build a large, high-volume data warehouse with Oracle on Unix using Informatica Corp. and MicroStrategy Inc. technologies. But the greater challenge in assembling any BI project team is identifying the necessary softer traits and skills that, with one exception, tend to be nontechnical in nature.

Here are seven traits to keep in mind when interviewing candidates for your BI project team:

1. Understand the Basics

Let's start with the exception: Everyone on the team, whether assigned to a traditionally technical role or not, needs to have good database skills. By this I mean that they can write their own SQL for testing and make their way comfortably around a data model, particularly dimensional models.

You can't expect every team member to be a database guru, but if members can't write a query, it's doubtful that they'll have meaningful roles on the project and play well with others. The degree to which this skill is necessary may vary with the team's size.

2. Know What You're Trying to Do

Understanding different approaches seems pretty fundamental, but you'd be surprised how many folks have data warehousing or decision-support systems (DSS) on their resumes but can't explain the difference between OLTP applications and DSS applications. One interviewee responded, "Well, the data model is a little different."

What an understatement! The differences between OLTP and DSS applications are huge, and if a team member approaches the BI application or data warehouse with an OLTP mindset, the project is doomed from the start. Everyone on the team ought to have an understanding of data warehousing terms and concepts. Key topics that should come up are:

• Subject orientation vs. business function orientation (for example, customer vs. accounts receivable)
• Time variance
• Volatility or lack thereof.

The wrong mindset is more common than you think: My team has assessed numerous data warehouses only to find the performance problem originated with the project DBA insisting on using OLTP techniques such as triggers, foreign key constraints, and excessive indexing. And the DBA couldn't understand why a load of 12-million rows took seven hours!

3. Focus on the Big Picture

Folks who frequently work on operational systems usually know one functional area fairly well but are unfamiliar with the other systems in the enterprise. Worse, they often don't know how their system fits in the overall business picture or how it contributes to achieving the company's strategic goals. People who can't see the big picture rarely make good BI team members for a number of reasons. First, BI projects, by their nature, involve data from a variety of business units that's used to measure and analyze across functional boundaries. A fundamental understanding of how data is used at an operational level and how it's combined with other data to provide insight is very important.

Second, a lot of decisions are involved in building a successful BI application. Knowing how the BI solution corresponds to corporate goals is vitally important. If the big picture isn't kept in mind throughout the process, it's very easy to make a decision that has detrimental effects further along in the project — or in subsequent projects.

Finally, successful BI projects can often be tied directly to one or more corporate strategic goals. Knowing this and ensuring that the resulting application supports the goal is the responsibility of each team member.

4. Stay Focused, Flexible, and Pragmatic

While keeping an eye on the big ball of wax, your team needs to be able to roll with the punches and adapt quickly and easily to change. I'm not advocating that your project be made up of "coding cowboys" and that you shouldn't use good project management and change control techniques. Your team needs to be able to mentally adapt to the changes and be able to incorporate them — while still keeping the company's and project's long-term goals in mind.

• Privacy
• Your California Privacy Rights
• Copyright © 2008 United Business Media LLC