Making data meaningful

Since the US subprime mortgage crisis started in 2006, more enterprises have realized the importance of data integration, according to Tony Young, CIO of data integration and quality solutions provider Informatica.

"While not a silver bullet, data integration is the important process of making data meaningful for decision making when it's transferred between departments or across enterprises," said Young. "The enterprise may extend data integration capability externally to connect with software-as-a-service (SaaS) providers like Salesforce.com or with supply chain partners."

He noted the problem of different enterprise functions having different definitions of the same subject area. For example, sales, finance and human resources may define 'customer' differently from one another. "Data integration allows these functions to agree on a standard definition of the subject area."

Young stated that data integration by itself is "not enough", as enterprises need to clean up non-useful data before transfer to another repository.

Data integration may also involve taking data from legacy systems and organizing it into a single database, such that the data is meaningful to all organization stakeholders in the current context.

Important for acquisitions

Young noted that data integration is important during acquisitions, as the acquiring company would not want to retain duplicate IT systems to ensure that all stakeholders have a single view of data presented.

"Creating a single view of data means that all information on a particular subject area resides in one rather than several locations," he explained. For example, a call center without a single view of customer data may have to re-direct enquiries from one staff to another because the requested information is scattered throughout different systems in the organization.

Data integration should involve people, process and technology. "CIOs should first ask themselves what the organization's core business is before deciding on appropriate technologies to leverage on," Young said.

Next, CIOs should ensure staff are trained in a common skill set needed to use the data integration solution.

Looking at process, integration competency centers (ICC) or teams are formed to facilitate data integration through standard data management policies. Depending on enterprise needs and policies, the ICC may comprise only of internal staff involved in data integration or a mix that includes expertise from the data integration solutions provider.

Young stressed the need for CIOs to have a data integration strategy and to roll it out step-by-step. "For example, data integration could start with only two or three departments before involving more functions and business partners."