Adopting an Information Integration Strategy

Wherever you look these days, you are bound to come across products that promise to provide the ability to integrate information from across the enterprise. Products designed to help companies solve this problem are often called Enterprise Information Integration (EII). EII is the technique used to bring various data elements together in an aggregated or federated method. It typically relies on caching data and access to the data. The actual data sources are left alone. The heart of many of these solutions is a meta data management capability that insures that the meaning of data across sources are consistent. Enterprise Information Integration is different from operational data stores or data warehouses because they are intended to help organizations aggregate information that exists in many different forms in order to attack a specific information management problem.

Every day there seems to be a new product introduced into this market. Products that include EII functionality that Hurwitz & Associates has analyzed include IBM’s DB2 Information Integrator, Avaki Enterprise Data Integration Framework, BEA’s Liquid Data, Composite Software’s Composite Views; Metamatrix Server, and Digital Harbor’s PiiE Platform. Some of these companies are pure play information integration platforms, while others provide information integration as part of a broader set of offerings.

Why are there suddenly so many products aimed at Enterprise Information Integration on the market? In essence, companies focused on this market are attempting to solve a critical problem: CIOs are now faced with a myriad of heterogeneous information silos that have been deployed throughout their organizations. These silos may contain both structured and unstructured information. Organizations need to be able to dynamically access and leverage this information, along with external information no matter where the information is coming from in order to be more competitive. For example, companies that are merging often must be able to link data sources from their respective organizations together in many different ways in order to move to an integrated company. Likewise, many companies are trying to figure out how to arrive at a single view of their customers across product lines and across their business partners’ offerings

Given the critical nature of the problem, it is clear that this issue will remain at the top of the list for corporations of all sizes and shapes for quite some time. It is important to understand what the issues are when you are thinking about information integration. Here is what I believe are the major criterion for adoption of an information integration strategy.

1. Build a strategy that assumes that the type of data, the structure of the data, the meaning of the data, and the scale of the data will change in an unpredictable manner.
2. A clear meta data infrastructure is imperative to creating a safe enterprise integration strategy. If definitions of data elements across a multitude of information sources are erroneous, management will be making decisions based on faulty data.
3. Ask yourself what you are going to use the data integration for, because this may lead you to consider a certain type of integration project. For example, there are times when a warehouse is the right approach when companies need to put a stable selection of data elements together to solve a very specific information need. Likewise, an operational data store is useful when a cohesive set of information needs to be supplied for ongoing decision making. Enterprise Information Integration is the right answer when the solution needs to be dynamic.
4. Context is one of the most important information integration issues. A data item from one data source must be able to match the business intent of a data item from another source. For example, two different data sources might have pricing information about a complex product line. In one case, the prices listed were based on a one-time special offer. If the integration does not treat that special pricing as an anomaly it could skew the results. I
5. It is not enough to get information into an integration platform – it is also important to have a consistent way to query the information.
6. Many organizations think about the information housed in their databases as the critical issue in data integration. However, the issue is a lot more complex. Some of a company’s most important information is in unstructured forms such as documents, email messages, customer support interactions, transactions, and information coming from packaged applications (ERP, CRM, and so on). Companies need to think about how they want to access and utilize this kind of information.
7. It is important to make sure that individuals within the organization understand the value of the new ways that information is being provided. People naturally like to stick with what they are familiar with. They will only change if they are convinced that the new approach to information access will make a difference in their effectiveness.

All of these issues are critical to implementing an enterprise information integration strategy. Even if an organization follows these guidelines, there is an overarching issue. Organizations must look at this holistically. Each industry has its specific issues that make information different from one market to the next. For example, retail organizations need a combination of customer information, and supplier information including significant information about buying patterns and trends. Pharmaceutical companies, on the other hand, need both information related to both the manufacture as well as the discovery of drugs. Drug discovery information is very different from the information needs of other markets. You get the picture. When looking for enterprise information integration solutions you must keep the complexities of your own industry in mind.

The bottom line is that enterprise information integration is an important platform for organizations needing to look at the relationship between a variety of information structures in context with solving a specific business problem in real time. Enterprise information integration offers the potential for management to leverage the vast amount of information to not only look for the obvious answers to questions but to be able to see trends and anticipate customer needs. Since EII is so tightly tied to corporate strategy, it is important to get the implementation right.

Judith Hurwitz is president of Hurwitz & Associates, a software strategy and research firm in Wellesley, Mass. She can be reached at jhurwitz@hurwitzassociates.com.