The IT industry is awash in a sea of data mart theory. This abundance of theory isn't really a problem unless your boss expects you to implement a data mart in the near future. If you're under that kind of pressure to deliver, you need clear, practical advice on how to organize and execute your project.
This month, we'll begin a three-part series designed to help IT managers who find themselves searching for practical ideas that will help them plan and build that first data mart. This series will take you step-by-step through the data mart development process and begins where all successful projects start--with sound planning.
The objective of planning is to take the ideas and intentions of a group of people who see the need for a data mart in their organization and convert them into a formal, planned, staffed and funded project. The goal of your planning efforts should be to insure that the project is firmly set up for success. You must do all you can to maximize the probability of producing a high quality data mart on budget and on schedule.
Define Goals and Scope of Project
The first step toward achieving this objective is clearly and explicitly defining the goals and scope of the project. To accomplish this, you must negotiate an agreement with the project sponsors on the subject area for the data mart. This will specify the extent of your star schema(s) and the source systems from which you will be loading data. Focus on goal statements that are concise, meaningful and achievable. Goals like: "improving the effectiveness of the company" should be avoided at all costs.
Once your goals and scope are firmly established, you must describe the approach you intend to use for the project and define a team structure which can effectively manage and carry out the work.
Outline Major Activities
When describing the approach, remember that your goal is to outline the major activities that will take place over the course of the project, the sponsor's involvement and, most importantly, when the sponsor will begin seeing results. Don't get bogged down in an overly technical methodology treatise. Provide a clear, concise, overview of the work to come and set the stage for defining the roles and responsibilities with the project team.
When it comes to project teams for data marts, smaller is better. If your goal is to deliver a production data mart within four to six months, a large unwieldy project team is a prescription for disaster. The optimal team size is somewhere between three and five experienced professionals. Once the team has been identified and assigned roles for the project, you are ready to move on and build the detailed project plan.
Project planning for data marts is no different than for any other software development effort. Like any other plan, you must include a detailed schedule of sequenced activities, assignment of activities to team members and a labor cost projection. Your plan should also include a list of activities that cover project management and administration. These activities should include such things as change control, progress reporting and cost oversight.
So far, we have not touched upon activities related to the selection of data mart tools or data quality. Neither of these activities should be part of your project. If a detailed evaluation of database, query, extract, transformation or load tools is required, it should be conducted as a separate project before trying to build your data mart.
This guideline also applies for any data quality audit or data clean-up effort. You should not, under any circumstances, try to embed data clean-up activities into your data mart plan. If you do, chances are your project will fail.
The planning phase of your data mart project as outlined here should take no longer than two weeks. If you find yourself spending substantially longer than this on planning, it may be an indication that you are over-engineering the plan. Remember, the key to data mart success is clarity and focus, and this guideline also holds true for your planning efforts.
Next month we'll pick up our discussion of the data mart development process and focus on prototyping.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access