Data warehouse is implemented in an organisation with the help of data architecture schema.  Elements which are specific to the company or organisation are defined in data architecture schema.  For instance, the administrative structure should be designed according to the real life undertakings of the company’s administrative department so that data resources can be managed to simulate the administrative department.

In addition, there should also be a database technology description of the methodologies to be used in storing and manipulating the defined data. A good design interface to map data in other systems is also a good way to implement a data warehouse along with infrastructure layout on how to support common data operations such as data imports, emergency procedures, data back ups and external data transfers.

When there is no proper guide or framework in the implementation of a data warehouse, common mistakes which could have been easily avoided will occur. A common pitfall would be different operations with data which makes it difficult to control the flow within the system. The result would be data fragmentation which can have tremendous negative impact such as potential increase cost. The problem with data fragmentation is typically encountered with rapidly growing business organizations or those having different lines of business in the one company.

A company trying to acquire other business or having mergers with another company may potentially experience difficulty with data fragmentation. For instance, in the field of manufacturing and retail, multiple order entries or order fulfillment systems may be poorly integrated resulting in data fragmented since stored at different storages in different locations. In another instance, poor integration with delivery services over multiple channels such as the web, retail offices and call centers can also result in data fragmentation.

The fundamental cause of data fragmentation also often lies in the complexity of an IT infrastructure especially if there is an absence of an integrated architectural foundation which is substantial in the interoperation of big volumes of heterogeneous data from various applications and business data accumulated over many years. It is not uncommon for business organizations to involve and have significant changes in business rules so an IT infrastructure should evolve as well, and this means the company has to invest more.

In many companies, more than 50 percent of the budget for IT operations is focused on building and maintaining points of integration especially when dealing with legacy systems dedicated to supply chain, finance, customer relations management and other mission critical aspects of the business.

Problems related to data fragmentation can be serious and relatively difficult and costly to attend. But they can be prevented with a good data architecture and IT infrastructure design that takes into consideration the future growth of a business organization.

The data architecture phase of an information system planning, when properly and carefully executed to the tiniest detail, can force a company to specify and draw a line between internal and external flow of information. A company should be keen in seeing patterns developed over the years and trends for the future.  From this particular stage, it could be highly possible that a company can already identify costly pitfalls and shortfalls related to information, disconnection between department and branches, and disconnection between current and future business endeavors. At this stage alone, more than half of the problem stemming data fragmentation is prevented.