This week’s episode of Conversations with Kip is a conversation with Jim Hladyshewsky about his first assignment with the SAFR team, called “Show me the Money”, what now might be called a data science method for financial data.
Show me the Money is a way of rapidly determining business rules and logic for posting processes without reading code. It recognizes that patterns of financial systems are quite systematic, and always determinant (people always want clarity about where the money is, where it came from, and where it is going.) These characteristics are shown in the way financial data is processed in financial systems.
We developed this approach on a legacy financial system, a General Ledger that was effectively highly integrated with a home-grown “ERP” system for a large property and casualty insurance company. Jim could have spent weeks trying to determine all the business rules that were applied to the data by reading code; but reading code is very slow.
Instead of doing that, Jim determined what the boundaries were for the system, locating all the input files. He then located key control reports and the journal entry outputs from the system. Knowing that all financial transactions must be accumulated into journal entries which are posted to the General Ledger gave him confidence that understanding how this key file was made would allow him to have a framework on which to hang all the associated business rules.
He then used simple prototyping approaches–what we today would call data science approaches with tools like Python and Spark–to try to guess at how to produce the output file and control reports from the input files at hand. Working with a business analyst and a technical person, they iterated over the code for a few weeks. When they found some problem, at times they would go read the code relative to that specific condition, until they had successfully produced a file for one day’s work.
They then attempted to do the same thing with a week, and later a month of data.
In the end, a very useful analysis of what the boundaries of the existing system that needed to be replaced where, and the rules that were used to produce the highly-trusted financial reports that had been used for a score of years. Making changes then was possible in implementation of the new system because the delta from the known was clearly understood.