Data warehousing has become increasingly important to modern companies as the amount of data necessary to remain competitive and make informed decisions about the future of a company has become greater and greater. A company's entire data warehouse often consists of a collection of data marts, sets of data that typically are focused on a particular department or area of the company. Combined into a data mart warehouse, these data marts help support the information system of the entire company.
A data mart might be dedicated to customer data or sales information, or it could be constructed specifically to support a company's human resources department. In some cases, data marts are compiled from data from the main data warehouse. Many companies, though, take a bottom-up approach to data warehousing and instead build the overall data warehouse by compiling data mart data into smaller systems, then combining them into a company-wide data warehouse.
One important factor in building a data mart is usability. Because the data mart usually is aimed at supplying information to a specific department for a specific purpose, the data must be easily accessed and processed. A company's Information Technology department or Management Information Systems specialists often focus on this area, ensuring that an easy-to-use interface exists so that employees, including managers and other high-level personnel, can access the data mart data and use it within the course of their jobs.
Individual data mart design depends on the specific needs of the company or the individual department. Most data is organized into relational databases, with a database management system implemented company-wide in order to let individual employees easily access data, generate reports or perform other necessary day-to-day functions. Some companies require more complex data processing that makes use of multidimensional databases. Still others use database functionality that allows multidimensional data analysis while still using the simpler and less costly relational database management system.
For the most part, data warehouses are built by adding data as it comes into the company's systems. Updates to data are not as important to an overall view of company performance as data that has been added to and compiled over a longer period of time. Managers can use data from the data mart to view company trends, sales patterns and efficiency of individual departments and to make forecasts about future performance. At this strategic planning level, the data mart and data warehouse are invaluable to those planning the future of the company.