In the first of a series of articles on MDM, Richard Skriletz of RCG starts by trying to define master data. RCG has an excellent reputation as a consulting company, but after saying some sensible things the article seems to me to get tangled up. To me master data is anything that is shared between multiple systems. This is captured well in the Wikipedia definition which Richard references, before immediately going on to try and “improve” on this very clear definition. He wants to distinguish between master data and “reference data”, in his case by splitting the world into physical entities like product and abstract entities like organisation. I have written about this before. Not only is there no need for this distinction, it can be misleading, and cause one to start treating data in different ways when there is no need to. I’m not sure whether IT people just love to classify things (all that being brought up on “int”, “char”, “varchar” etc) but this desire to split master data up is just confusing. Data that is shared between systems needs to be managed, and it needs to be managed whether it is physical or abstract. By imposing artificial distinctions between types of master data we introduce complication where none is necessary. William of Ockham had the right idea on this, and he did so in the 14th Century.