Problems: ETL of Boarding Loans
ETL of loan boarding is a complex data integration process that is a significant pain point for the mortgage servicing industry. Mortgage transfers contain a lot of files, requiring a lot of data mappings to be managed. Compound that with every mortgage servicer have slightly different file format standards so the loan boarding team need to maintain a history of mappings for every prior servicer. Some of these files can be in poorly formatted files like the T309 which requires specialized or custom text file parsing functionality that not all ETL tools can handle. Another area of complexity is the significant amount of validation required for financial and legal compliance. Validation can be split between data validation and business validation and each require a separate technology. Data validation typically is implemented during the ETL process of loading the files into the staging tables. Business validation after the data is loaded into a staging area and has passed the data validation. Business validation is more or less auditing the loan history. Nonperforming loans’ transactional history as even more mappings and validations.
The mortgage industry’s complex database schema requirements can leverage some of the newest ETL and application technologies. The combination data rules engines, metadata management and json processing provide significant automation. ETL technology cannot solve all issues so business analysts will still need to intervene. The loan boarding application should be able to manage the analysts’ tasks with a modern, low code workflow management application. The loan boarding process should be able to quickly complete a “gap analysis” during the ETL process and after the data is staged data. This should in turn create a “to do” list for the analysists with predefined workflows. The Semiotic: Semantic ETL Tools provides the key ETL technologies to meet these challenges. Combined with one of the many workflow managements system including open source. A semi-automated load boarding system is attainable.
With all the data mappings to maintained using metadata/json it is much easier to manage then a code based solution. Metadata allows a clean “gap analysis” this is not available to manual coding. Metadata allows fixes to be implemented by business analyst verses the expense of maintaining a ETL programming staff.