In the early days of computing data processing was riddled with incompatibilities. Most manufacturers used proprietary operating systems and file structures. There were many obstacles to be overcome to effect even the simplest transfer of data between applications. Now, thanks to Microsoft, we think nothing of copying and pasting a table of numbers from an Access database to an Excel spreadsheet or from an accounting system to a formatted Excel worksheet. The standardisation of operating environments and databases coupled with the convergence of computing and telecommunications means that we often take for granted the ease with which data can be moved between applications. So why is it still so difficult to merge, say, sales data from an ERP system with a data warehouse and a forecasting system? The answer is often because of a clash of metadata. Gary Simon, managing editor, FSN explains.
Physically transferring data between systems has never been easier. Data storage devices have become smaller and more portable whilst their capacity has grown exponentially. Shifting data, unless you are a bank with millions of transactions a month is rarely an issue. The real issue is in understanding the meaning of the data that has been transferred. The so called metadata data about data- describes the structure of data that is held within an application. In an accounting context common examples of metadata include, the chart of accounts, the organisational hierarchy of cost centres and divisions or perhaps product groupings and accounting periods. These descriptions form the backbone of data and effectively give it meaning. The trouble is that metadata is proliferated across most organisations despite the best endeavours of IT departments because applications grow like topsy. For example, the group finance department acquires a consolidation system and sets up a hierarchy of divisions and cost centres which is similar but not the same as the hierarchy held in the ERP system. The marketing department buys a merchandise planning application that has different product groupings from the general ledger and the HR department uses employee grade structures that are different from the payroll system. Essentially, everyone in the organisation is talking in slightly different terms about the same thing ? the metadata is different. So, in our example, when someone wants to find out the direct costs of labour in the marketing of a particular product group they cannot do it without resorting to working in a number of off-line spreadsheets. Even if the metadata could be brought into alignment for an instant in time, someone can change a structure, add a department or a new staff grade and the whole system can be thrown out of equilibrium again. Differences in the way the data itself is held just adds to life s rich tapestry of problems. For example, an employee code could be eight characters long in one application and seven in another. Failure to synchronise metadata and data can have serious consequences. The inability to produce reports as required can hamper the delivery of new and important business insights and considerable effort involving highly skilled finance professionals can be wasted trying to fashion improvised reports from data manually re-keyed into spreadsheets. In addition, there is the added risk of introducing error or failing compliance thresholds. If differences in data and metadata are responsible for holding back progress in operational and performance reporting, what can be done about it? Fortunately, the problems of dysfunctional metadata are being taken very seriously by many software developers. A number of specialised metadata management tools, such as Metadata Management from ASG are available which allow metadata to be defined retrospectively and managed in a separate data repository. But they have to deal with enormous complexity. Many legacy systems are comprised of packaged ERP systems and countless bespoke or tailored systems. Defining and managing metadata over a sprawling infrastructure like this can represent a Herculean task. In the business performance management BPM space suppliers see the management of metadata as the key to securing an effective performance management regime. Organisations such as Hyperion and SAS, Outlooksoft, and Cartesis have invested millions of dollars in developing a broad range of applications that share the same metadata foundation so that finance users can produce reports from budgeting, scorecarding, and consolidation systems that use the same data definitions, business rules and metadata structures. In many cases, the way in which metadata is managed allows data from external sources to be brought into the same environment permitting a complete picture of management information to be maintained for reports and analysis. With the increasing complexity of reporting and compliance, finance functions simply cannot afford to waste time on data management. Therefore metadata management must be a primary consideration for the finance function before it buys a business intelligence solution or performance management suite. You may be unaccustomed to the technicalities of metadata management but ignoring this critical area could be costly in the long run. Source: www.fsn.co.uka>