Over the past decade data warehouses have not only become strategically important for big finance institutions, but they are also indispensable for a broad range of companies.

The question is no longer whether or not to have a data warehouse, but whether a data warehouse is truly beneficial for its users in business and marketing, risk management, controlling and corporate management.

Common problems of data warehouses

Data warehouses are among the most complex systems used by companies. They are connected to key production systems and integrate within themselves sophisticated processes of data extraction, transformation, cleaning, storage, aggregation, analysis and distribution. At the end of this complex chain of processes, data in the warehouse are used to support operational and strategic activities and decision-making.

A less than optimal initial design of the data structure, progressive growth of data, rising numbers of users and their queries to the data warehouse, implementation of new functional improvements and numerous other factors cause the gradual growth of the data warehouse’s operating costs. Other problems that can occur are, for example, gradual reduction of the number of requests that are realised, longer period required for implementing change requests, declining quality of outputs from the data warehouse, inconsistent or incorrect information in reports and so on.

How are these problems solved?

Our decade of experience gained from projects involving expansive data warehouses indicates that most problems observed by business users are associated with the data warehouse’s architecture and how the following technical aspects are handled:

• ETL – process of extraction, transformation and loading of data from individual systems into the data warehouse
• Metadata management – method and means by which data, their structure and mutual relationships are described
• Data quality – cleaning and standardisation of data
• Reporting – transformation of data into information that assists users in decision-making
• Software process – the extent to which the data warehouse is approach as a standard project of development, operation and expansion of a software system