An important aspect of implementing a Master Data Management solution is the unique and uniform recording of relationship data. Since there are multiple applications in which (product-oriented) work is performed and in which the same relationships are often re-stored, it is not guaranteed that those data are the same.
When a particular application is connected to the MDM system, that data is validated and standardised. This leads to an increase in data quality. Equally important, error-tolerant and high-precision matching is performed when data is incorporated into the MDM system.
High-precision matching takes advantage of the aforementioned fault-tolerance, but at the same time sets very precise rules to determine which matches are relevant and which are not. By taking into account the general similarity as well as the similarity of specific elements, it can be very accurately determined to what extent two relationships are a match.
Does only the name differ (e.g. "Matijssen" versus "Mathijssen") or is the date of birth also slightly different (e.g. 13-02-1955 and 16-02-1955) or is even, despite similar initials, the first name too different (e.g. "Joan" versus "Johan")?
High-precision matching therefore leads to fewer false positives (wrong matches) and prevents false negatives (missed matches). This ensures that in the MDM system only the correct records are processed into one Golden Record. For the Data Steward, this leaves a limited set of potential matches to be assessed.
An important aspect where an MDM system adds value is in the area of duplicates and data quality. While processing data from source systems, validation and standardisation of that data takes place. On the one hand, the aim is to record data in a uniform manner, but of course you also want the best quality in your Golden Records. That's where validation comes in. Data is provided with an indicator that shows the quality of the various components and with this, decisions on compiling the Golden Record can be guided, but also a Data Steward can take targeted actions.
Something similar applies to matching data from the source systems, where duplicates are flagged. Above a self-defined limit, those customers will be considered certain duplicates and automatically merged. The group below that concerns the possible duplicates and a Data Steward can then do an assessment on those and decide whether or not to merge those records.
An active Data Steward ensures that the quality of data in the source systems keeps increasing. Through continuous synchronisation, data is improved and by assessing duplicates, duplicate clients in the source systems can also be merged.