Match & Merge
In the today’s competitive environment, organizations need information to be as complex and complete as possible to make the right decisions quickly and serve their clients in the best possible way. With data distributed in many different systems, various data integration activities come into the spotlight. The siloed approach is being replaced with building centralized data storages to serve all business units and users across the organization.
These centralized data storages, such as data warehouses or customer ODS, require proper data consolidation to provide desired information about the key data entities. Without it, it is difficult to properly answer questions like: "How many customers do we have?," "Which product does this customer have across the whole organization?," and many others. Organizations simply need to have a single view of the customer at their fingertips.
The goal of the Match & Merge process is to create groups of records that represent (with the highest possible probability) one real entity and to provide consolidated information about the entity. Match & Merge processes are the most important step in building a single view of the customer (or any other master entity).
Business departments (especially Risk Management and Marketing) usually have different viewpoints on what constitutes the same entity. Ataccama supports this by its rule-based approach, which enables setting the certainty levels according to business requirements.
The typical scenarios involve the following:
- Matching. Identify records from various sources which represent the same entity (usually Customer, Patient, Product, Location, etc.)
- Merge [Golden Records]. Create representative master entity records, once again, rule-based. Typically the Golden Records are either selected or composed of the individual attributes with required characteristics (e.g., highest quality, latest update, source system preference)
- Deduplication. Identify duplicate records in the data set, select survivor records
- Identification. Distinguish new and existing records entering your systems, avoid creation of duplicates in your Master Database. Identification is a modified "front-end oriented" matching process which decides whether the incoming records belong to pre-existing customer groups or constitute a new one
- Relations identification. Create relations between entities for various marketing and risk management processes, such as Householding and Commercial Householding, identification of co-debtors, etc.
The Match & Merge processes shall follow the Data Cleansing activities that segment the data into standardized formats and validate it, in order to achieve better and more reliable results.
Match & Merge processes represent a crucial component of any master data management solution.
Ataccama supports the solution
Ataccama provides both the technology (specialized algorithms, predefined Match & Merge plans ready to be adapted to customer requirements) and also the business logic (knowledge bases, country-specific cleansing and standardization rules, and the best practices in this area).
The technology is designed to handle Match & Merge in both online and batch mode, and incremental data processing is a key functionality.
- Elimination of duplicates inside and across systems/databases
- Single view of customer (or supplier, employee, patient, vehicle)
- Identification of data inconsistencies within groups of records
- Identity resolution support (complex approximative matching)
- Rule-based Golden Record creation and data enrichment
- Persistent identity (incremental processing)