Data Warehouse Project Management
Introduction
Building a data warehouse usually involves more human and capital resources than any other information technology project an organization has undertaken. Because of the magnitude of the project and
its impact on the organization, the potential for disaster is great. Many data warehousing projects stall because of inefficient technical infrastructures, lack of executive-level support, and
inexperienced project teams, and long-term commitment.
As a data warehousing project manager, there are some proactive steps you can take to build a solid foundation for the warehouse so that your project not only proceeds smoothly, but actually has a
positive impact on the organization. The foundation for the data warehouse is built on the quality of the technical infrastructure, executive sponsorship, information systems (IS) staff, and the
long-term commitment of the team.
Examine the Existing Technical Infrastructure
Before you build a multi-gigabyte, two-million dollar data warehouse, be sure your organization has-or can build-the appropriate technical infrastructure for such a massive system. An inefficient
or undersized infrastructure will impede data access and render the warehouse unusable.
Robustness is of primary importance in the technical infrastructure. Robustness means keeping all the LAN components running as smoothly as possible on a regular basis. If your LAN/WAN system does
not function well regularly, then your warehouse-and any other systems that rely on this structure for communication and data retrieval-will not either.
Data access efficiency is also a critical infrastructure issue. Users want to have almost-immediate response to their requests for information. In order to maximize access efficiency, maximize the
amount of data that can be transmitted while minimizing the time it takes to transmit it.
One way to increase access efficiency to build data marts connected to the main warehouse. In a data mart, each facility in the organization has a small server that contains a subset of the
information that is stored in the warehouse. This small server is updated with warehouse information on a regular basis. When each facility’s users run queries, they run the queries against the
small server, so they gain the benefits of their quick LAN without having to sacrifice performance.
Secure Executive Sponsorship
The next step in warehouse construction is to secure committed, long-term executive sponsorship for the life of the project. The role of the executive sponsor is to shepherd the project through the
organization throughout its lifecycle. Most importantly, the executive sponsor must foster a culture of communication, direction, teamwork, and excitement among the many groups that must work
together to build the warehouse.
The executive sponsor must also have enough power to direct department managers to free up team members for the project, at least on a part-time basis, for the duration of the project. Most people
are overworked in the best of circumstances, and it is easy for them to put aside the warehouse project for more urgent, immediate tasks. The executive sponsor can alleviate confusion over project
priorities, however, by continually supporting the project and monitoring the performance of team members.
Assess the Caliber of Your IS Staff
Once you’ve secured the project’s executive sponsor, you can start building your team. The experience, quality, and commitment of the project team members will be vital to its success. How your
project team handles the day-to-day, labor-intensive work, as well as the unexpected challenges that will inevitably arise, will ultimately determine the quality of the final product.
The data warehouse technical infrastructure will likely be comprised of best-of-breed hardware and software components; therefore it helps to have people on the project who have a well-rounded
knowledge of many different information systems (IS) products and methodologies. The warehouse should also be designed so that speed is maximized and maintenance and updates are minimized. Project
team members who have eclectic IS knowledge are invaluable because they will likely have experience with many of the tools used in warehouse construction and be able to optimize performance.
In addition, subject-matter (SMEs) and legacy-system experts who are knowledgeable about the data and the systems that will feed the warehouse are indispensable additions to the project team. These
SMEs and legacy systems experts will provide valuable background knowledge of what correct data is supposed to look like and where it’s supposed to come from.
Less important-but still helpful-are project team members who have experience in data warehousing. Realistically, however, there are few IS professionals who have warehousing experience. Data
warehousing experience is not critical, though. It’s much easier to build a strong internal team that knows your IS environment, and then teach that group about data warehousing. It would be a
much harder effort-not to mention more expensive due to their high price-tag-to teach data warehousing experts about the particulars of your IS environment.
Demand Commitment for the Project
In addition to being knowledgeable, the team must also be committed to delivering a high-quality product. The payoff is a long way down the road from the project’s inception; that can be
discouraging for team members. Project team members must understand that this payoff is worth the wait. They must be committed to the warehouse and not be lured away from the project by other
projects with shorter-term payoffs.
Project management must also support the resources to the project and not assign warehouse team members to other large projects. If resources are continually pulled away from the warehouse project,
team members will begin to question the warehouse’s importance to the organization. They will become discouraged about being assigned to a project that is not a top priority.
Develop a Realistic Project Plan
Once your team is gelled, you can begin your detailed planning process. Planning is not a job for IS troubleshooters. It requires thorough, long-term thinking. Avoid letting the project derail
because of impossible-to-deliver-on or insufficiently-defined expectations. The project plan, not promises made to funding parties and user groups, must drive the development of the data warehouse.
This does not mean that the project plan must define, up front, the exact legacy systems and data elements to be contained in the warehouse. It would be almost impossible to define the data
warehouse with such detail at the outset. Instead, a well-drawn-up project plan should be at a high level only and consist of an incremental implementation by data type to be populated, receiving
department or user type, and maintenance responsibilities.
For the Future. . .
Having a solid technical infrastructure, secure executive sponsorship, knowledgeable and committed team members, and a thorough project plan is a foundation for success, but it’s not enough.
Warehouse project managers must play a dual role for the life of the project. They must continually shore up the foundation they’ve built while looking for ways to improve the quality of the
product-even while it’s still in development. Project Managers must also look in the crystal ball of the unfinished warehouse to predict and plan for its future.
Good project management alone cannot guarantee that the warehouse will meet everyone’s expectations, but it can increase the chance for success exponentially. Building a warehouse is an arduous,
long-term task. It takes a combination of forethought, patience, and skill to bring a warehouse project to successful completion. Thankfully, the payoff of vastly improved data access is well worth
the trouble, and the wait.
© Copyright 1996 – Systems Techniques, Inc.
Stacey Herdlein is a Project Manager at Atlanta-based Systems Techniques, Inc., an information technology firm specializing in delivering value-based, integrated information to the
enterprise.
