This green paper presents SysCore's current thinking on the practical approach to defining and using Reference Data, especially within financial organizations. (Practical green papers share SysCore's experience; white papers are more visionary.)
SysCore uses a relatively broad definition of reference data:
Reference data is any kind of data that is stable and reusable across business lines and applications. It is often data whose primary purpose is to categorize other data and can be used to define transactions.
Reference Data is typically 'referred to' by other data within the organization (especially transactions). The best examples of reference data are those that don't change much, as the reference is always valid.
This green paper discusses what reference data is, describes different means of classifying reference data, enumerates the different types of reference data and examines some of the implications of these different states. The document also examines how reference data may be classified based on how and where it is created and maintained and how that affects its system of record (SOR) status.
The paper describes the vision of how reference data should be managed and discusses the policies, procedures, roles and responsibilities that are essential to establishing a sound reference data management foundation.
request full paper
This paper was developed based on a SysCore solution delivered at one of the largest Financial Services Companies in the world. The solution integrated global post trade data for customer delivery via the web or via a customer services representatives from nine large back office legacy applications.
There were three issues facing this firm:
- fragmented data
- inconsistent data
- the technology in use was inflexible and didn't support Web access.
Based on analysis of the situation, SysCore came up with 7 recommendations:
1. Don't do all of the project at once.
2. Develop a core data model from the bottom up.
3. Capture the business data in real-time (without being rigid).
4. Work towards more consistent data without expecting perfection.
5. Accept only a single version of any incoming message.
6. Design for loose coupling between parts of the application.
7. Design for warnings of data inconsistency where it can't be removed.
The benefits of the system developed jointly by SysCore and this firm, following these recommendations, were improved customer service and clear competitive advantage. The project is continuing, with other future benefits described in the paper.
request full paper |
SysCore is constantly updating its library of green papers. Check back often.
Contact Us to request complete green papers.