Information Management Enlightenment

Gartner have recently been using the term "enterprise information management" as a blanket term to describe the technology and processes around a company's efforts to control and best use its information assets. The term extends beyond structured data into text, and even to digital content such as movies or music. As they identify, a key to making progress in such a potentially monumental task is to resolve semantic inconsistencies across the technical boundaries. The problem will be familiar to anyone who has worked in a large company. A product code used in the ERP system has a different code in the CRM system, and a different one again in the manufacturing system. There are godo reaons why such differences have emerged. If we take a physical product, then a manufacturing group will care about the materials that go into that product, its manufacturing process, and perhaps health and safety information associated with it. From a distribution perspective, its dimensions are important e.g. how many will fit in a container. From a marketing viewpoint it is important to understand the branding used, the packaging (perhaps the product is marketed in different ways in different countries) and the pricing. Each business unit cares about certain aspects of the product, but has limited direct interest in other aspects, so it is hardly surprisng that the ways in which the product is classified are different depending on whether you have a manufacturing, distribution or marketing viewpoint. For example, even something familiar like a Big Mac actually has quite different recipes in different countries; in some cases it is no longer even made from beef (e.g. in India, where the cow is sacred). A branded automative lubricant will have quite different technical specifications if you buy a can of it in a hot country like Vietnam than if you buy apparently the same product in, say, Iceland.

These different perspectives cause a complex web of differing classifications and semantics to occur for products in a large enterprise, and it is similar story for other terms like "customer", where again the key points of interest varies dramatically if you are a salesman from if you are trying to deliver a consignment to that customer, to whether you are trying to collect a payment from them. This is not just of academic interest: according to a Reuters survey, 30% of operational errors (e.g. incorrect deliveries) can be traced back to poor quality data. Those who have ever had the joy of trying to change your address on your bank or savings account will be familiar with the issue that your details do not just occur in one system only in many banks.

Trying to manage the various types of data in a large company is a mammoth task, and one which is on an uncertain footing since brands and customer details do not stay constant forever. This underlying patern of change means that initiative which seek to standardize (say) product codes "once and for all" are doomed to failure, because the things themselves are changing. However progress is possible. The first stage is to gain an understanding of the data across the company, then to describe the processes used to update this master data, and finally to bring automation to these processes. There is no single technology silver bullet since business processes are just as important as integration technology, but there a number of technologies do help matters e.g. data quality tools to help identify issues, emerging master data management products to assist with process automation, data warehouse technology to help understand and classify reference data, and EAI technology to actually link up and automate processes once they are under control. "Think big, start small" is the mantra, starting with a manageable scope and going through this process: identify data -> capture the processes -> automate the processes. Modern technologies that are better able to deal with change, along with universal access to the internet and so to applications that can automate workflow, are starting to make it possible to begin this enterprise information management journey. Confuscious said: "A journey of a thousand miles begins with a single step", and companies can set about this journey with more confidence than ever before.