Data warehousing best practices: Part I

These data warehousing best practices will facilitate an organization’s development of an efficient data warehouse.

Data warehousing and reporting have always existed in their simplest form of management information system (MIS) reports. In the past, data warehousing was carried out more from the perspective of utilizing historical data. However, with the competition intensifying amongst vendors and with evolution of customer needs, data warehousing has grown to be a discipline.

This first part of a two-part series on data warehousing best practices focuses on broad, policy-level aspects to be followed while developing a data warehouse (DW) system.


Best practice 1: Ensure support and sponsorship from the CEO’s desk

As a data warehousing best practice, while considering investments, ensure executive buy-in. The sponsor of the data warehousing project plays a key role and it’s desirable that the CEO undertakes it. If the IT department propagates the data warehousing initiative, it will be viewed as an IT project and business users may lose interest in it.

Consider data warehousing as an organization-wide project and not a departmental one. The initiative may start on a departmental level for trial and error iteration, but gradually, it has to span across the enterprise.

Best practice 2: Develop a holistic RoI model for DW investments

The data warehousing project planning should identify the key performance indicators (KPIs) to measure aspects like revenue increase and operational ease. These factors should be quantified to the best of one’s ability.

A major bottleneck for data warehousing is that these projects are looked upon as additional costs. They should rather be viewed as enablers to improve the top line, bottom line, and operational efficiency.

Best practice 3: Consolidate historical data by beginning data warehousing early

As a data warehousing best practice, begin investing as soon as the organization implements complex operational systems like enterprise resource planning or customer relationship management. These systems are fertile grounds for data generation. Consolidating the data from these sources will immediately aid in establishing and maintaining good data quality. The consolidated data will fuel faster and more accurate business intelligence (BI).

Best practice 4: Know the domain

The DW will reflect how different businesses interact and interface with each other to meet enterprise-wide objectives. To illustrate, if a bank sets out to develop a DW, banking-specific domain expertise will be needed to customize the solution.

Best practice 5: Know the source systems well

Data warehousing is about integration. Another consideration is to find if the functional and technical nuances of the source systems interact collectively or work individually. For instance, a bank will have a core banking system, a credit card system, a PRA (purchase and resale agreement) system, and others. The method in which these systems generate data varies. Having knowledge to incorporate these variations in data warehousing is important.


About the author: Amit Agarwal is a seasoned BI professional, heading the APAC and MENA businesses at iCreate, a specialist provider of packaged data warehousing and analytical solutions to the global banking industry. 

(As told to Sharon D’Souza)



Content Continues Below

Read more on Data warehousing