Specialisation needed for data warehousing success

Susan Andre1999

[Johannesburg, 3 December 1999] – Data warehousing is not cheap but it can be very cost effective. The project manager who combines the relevant technical experience with authoritative management skills, coupled to an appropriate touch of diplomacy can keep the project's view firmly on the goalposts. This ensures cost effective delivery, justifying the expenditure and guaranteeing future returns on investment.

The critical success factors of a successful data warehousing project have unique aspects, which are largely due to the strategic nature of data warehouse deliverables and the complexity of data warehouse design and construction. To ensure those deliverables, the project manager assists the business in defining the information needed, identifying raw data sources, interpreting these into data storage designs and devising the often complex equations for data analysis.
The project manager has to be skilled in dealing with:

  • Business people who are often specialists in their areas and can be frustrated and impatient of the analysis and specification process. Their involvement is essential to ensure that the end result produces the strategically important information being sought.
  • The IT staff who look after the operational systems.
  • External service providers. The project manager must be able to advise on skills and manage the relationship.
  • Skilled (and scarce) data warehousing specialists.
  • Project ownership issues. This project must be owned and driven at executive level. Some IT departments may not appreciate this approach. In addition, there may be differences at executive level which could impinge on the project's success.

A well-managed and goal-orientated data warehousing project has the following characteristics:

  • Measurable progress. An iterative approach provides for carefully defined deliverables at regular intervals of six weeks to 12 months. The deliverables must be distinctly usable and not rely on further iterations (although they may be enhanced).
  • The full-time involvement in the project of a senior and experienced businessperson who has management contacts and influence. The project manager must maintain this relationship, without relinquishing responsibility or authority.
  • Understanding of the technical constraints of data warehouse design. The project manager has to be responsible for the quality assurance of the technical design. Ensure a robust and efficient infrastructure, which will deliver the information fast.
  • Managed time and skills of multi-skilled data warehousing specialists. Overlapping phases and iterations need carefully scheduling.
  • Knowledge transfer. This is vital. The design constraints alone mean that the project manager must ensure that once the project is finished, there are people left behind who know how to interpret a business requirement into a design, as well as knowing the intricacies of performance tuning the data warehouse. In addition, internal training and marketing has to be done to ensure that there is buy-in to the data warehouse and that it is used by decision makers.

The lack of IT management skills in South Africa has resulted in the development of the vanilla flavoured project manager. Successful data warehouse implementation relies on more than the ability to manage staff, technology and client expectations. Data warehousing needs project managers who are specialists and generalists. They have to be specialists in understanding the unique aspects of design and implementation, and generalists in a wide range of skills, both business and technical that are applied to a data warehouse.
Sandri Jordaan is the consulting services director of Sagent SA, the data warehousing company. She is a specialist project manager having been involved in data warehousing as both a technical and a business consultant.