Master data management is an essential element in IT governance and data assurance. Increasingly organisations now keep their master data online using databases and business platforms.
E RADAR's Will Roebuck looks at some of the management and operational challenges companies are facing when using master data across their electronic supply chains.
Products for sale are identified by their descriptions and key characteristics. They are also given a classification so that customers and supply chain partner can identify them easily. All such information relating to the product should be held in structured master data files that are accessed via ID numbers or Auto Identities. These are then linked to product and price master data files in an online database or platform.
This is known as master data management.
Other related forms of master data include technical specifications/design, product life cycle data and financial data (prices/costs). Master data are semi-static in that they have a time structure which defines when they are applicable; now, past and future.
Of course, setting up any database or data assurance platform also requires organisations to either own the relevant software or have a licence to use it. Contracts should be in place with software designers, developers, and platform operators to ensure the no-one's rights are breach. Value in the database is protected through Intellectual Property law.
Master data are most often relatively stable and are usually best sourced from the original owner. For example, product item data, including the technical specifications, should come from the original manufacturer. Each supply (and demand) chain participant should synchronise relevant data with their collaborative partners in advance of trading or communicating with each other. This can be done by the electronic data interchange of files, by electronic catalogues or via e-exchanges.
It is too easy to allow master data to become non-aligned both between and within supply chain participants. Strict controls are necessary for data maintenance and cleansing, for broadcasting new versions and for synchronisation. The difficulties of achieving common, accurate master data should not be underestimated, but neither should the benefits. There have been many instances of companies’ performance declining rapidly when master data relating to what has been produced or delivered are “out of sync” with what is supposedly in stock and/or being sold. Poor master data rapidly corrupt dynamic data.
Accuracy of master data
Experience in successfully implementing Enterprise Resource Planning (ERPs), such as SAP, demands that supply chain partners first get their master data correct. Too many organisations have great varieties of master data held all over the place, covering products for sale, bought-in materials, customers, suppliers, etc. The secret of success is to centralise and integrate these into one set of master data and to undertake any devolution of control with experience. Getting master data correct for developments such as ERP implementation also provides a sound basis for e-business across the supply chain.
Master data will vary to some degree according to the type of supply chain. The simplest Master data relate to the buyer, the seller, and the item being traded.
Types of master data
A typical health sector supply chain, for example, would have master files on: patients; diagnoses; treatments; members of staff, including their roles; locations; items and services employed; suppliers and manufacturers, plus their locations; authorities (Department of Health, safety bodies, clinical bodies, etc); agents (transporters, banks, etc).
Early auto-ID of master data
Medical members of staff give themselves an Auto ID as well as their location and the patient before undertaking diagnoses and administering treatment. This check would help to ensure that medical staff were treating the correct patient and to avoid later negligent claims for mistreatment. Medical symptoms would be measured and fed in to diagnoses master files with associated decision support systems. The number of measurements would be increased until there was an acceptably high probability that the diagnosis was accurate. From the diagnosis would come the treatment, which would also have to be cross-referred to the patient and to the staff member to ensure that these were appropriate. When treatment was taking place, the patient, location, staff member and item (be it a drug, set of surgical equipment, etc) would be cross-referred via Auto ID and master files to ensure accuracy, safety, accountability and traceability.
The correct items should be made available to the appropriate locations, at the right time, at the minimum total cost. Hence the use of items in treatments at locations needs to link back securely and promptly to those managing, ordering, transporting, supplying and manufacturing the required items. The successful employment of modern value chain management techniques, such as continuous replenishment, co-managed inventory, collaborative forecasting, planning and replenishment, collaborative event management, and tracking and tracing, depends on global standards for messages, data elements, master data, identities and Auto ID for the cost-effective use of e-business in their support.
An important part of master data management is item classification. For example, buyers may wish to know who the suppliers of particular types of products are. Managers may wish to know what the trends in each type of spend are. Classification or categorisation of items aids these types of analysis. Classification is sometimes confused with Identification. Identification is far more precise and is essential for modern supply chain management as well as for safety and security. Classification provides broad pictures of activity and NOT the close control needed for supply chain management and for safety and security purposes.
Classification needs to be done carefully so that the needs of all key parties are met. For example, the data analyses needed by a buyer may well not coincide with those of a clinician or engineer, nor with those of a senior manager or finance director.
Classification is in essence another set of data elements within a master file which help to describe an Item, NOT to identify it. Classification may be undertaken from many valid viewpoints – market or category segment; material use; technical characteristic; personal characteristic or role; etc.
Locations, individuals, assets, and all master data, may all be classified.
Master data are important for defining information in the past and future tenses as well as in the present tense, i.e. “as was” and “will become” data as well as “as is”. This enables such questions to be answered as: “when did (or when will) these prices apply?”; “when were these components in use?”; “when was the name changed?”