Johannesburg, 15 Aug 2012
With any enterprise project, it makes sense to follow a phased approach, pick the low-hanging fruit, deliver some business benefits early on, gain buy-in and support, then continue along the chain. It highlights the considerations of complexity and cost. No company wants to tackle projects that are highly complex and very costly. The opportunity for failure is just too high and the cost can be extreme. Businesses want to get stuck into projects that are low-cost and which have low complexity to give themselves the best opportunity for success and to get the most immediate and cost-effective benefits possible, says Johann van der Walt, MDM Practice Manager at Knowledge Integration Dynamics.
That's why I believe there is still a great deal of education to be conducted around master data management (MDM). In many cases, South African businesses are going directly to the high-complexity, high-cost end of the scale, and it's not because they're suckers for punishment; it's because they lack the fundamental understanding of the animal that is MDM.
Overall, MDM aims to provide quality information that business executives can use to devise strategies and make operational business decisions. The need usually stems from segmentation of business lines that results in poor understanding of customers, products, materials and a host of similar business components. There are multiple, overlapping and duplicated data entities, and that clouds the view executives have of the business, disabling them from making sound decisions about the organisation's future. That situation arises for a number of reasons. Businesses may have been run in a federated style, growing organically and allowed to operate independently, or there may have been mergers or acquisitions. It's a common scenario in South Africa, and, as you can imagine, there are a host of benefits to be had from clearing the fog and allowing the rays of reliable data to filter through.
First off, there are two strains of MDM: analytical and operational. On the cost/complexity scale, analytical is low-cost, low-complexity, and operational is high-cost, high-complexity. But they also have two very different aims. Analytical MDM, in 99% of cases, supports business intelligence (BI), ensuring that execs get good-quality information that they can use when formulating strategies and directing the course of business. Operational MDM is primarily aimed at achieving operational efficiencies, paring down processes, centralising data control and driving down costs.
In terms of a business' MDM maturity, the best place to start an MDM project is going to be with the analytical version. It creates an MDM foundation that demonstrates tangible business benefit without disrupting operations and requiring enormous change management that permeates the organisation through every operational touch-point. Spend a short time considering the ramifications of that and you begin to see where the expense and complexity begins.
Analytical MDM, on the other hand, doesn't necessarily delve into the operational processes, requiring that they be redesigned to fit the bill for central control over data. What it does is gain a high-level view of the data that exists in the organisation's dispersed and otherwise disparate systems and it then establishes reliable data to provide executives. The methods by which it does so are highly complex, but the upshot is that it knows which data is to be trusted, retained, maintained and promoted as the correct one to be promoted wherever it is required, regardless of the geography or business function.
It doesn't necessarily retain a central, primary source of all data throughout the business. It leaves that up to the established domains. But it knows where the data is and where there may be duplicates; for example, it knows which version is to be trusted. Thereafter, the MDM team can set about eliminating the duplication and all of the other data issues that may be apparent.
Share