What happened to a major financial service provider when they prepared their quarterly statement makes the perfect case for Beta 91: According to press sources, a computer error caused faulty bookings affecting hundreds of accounts, in some cases resulting in negative balances in the range of tens of throusands of dollars.
No doubt, discovering these balances will have been quite a scare for the customers. While the bank insisted that only a few accounts were involved, apologized to its customers and quickly corrected the postings, still, such an event can cause long-term damage to the corporate image.
Implement Measures to Prevent IT Errors
When we read about such incidents in the press, we are left clueless as to what actually caused the problem. A possible cause could be insufficient data quality assurance or an inadequate comparison of target and actual data. Beta Systems develops specialized software tools that enable the structured and convenient definition of automatic reconciliation routines. They help companies perform these checks and thus prevent errors that can cause widespread damage. The tools integrate with existing processes and can be launched directly from within the program code.
Immediate Response and Troubleshooting
After comparing the control parameters and identifying possible errors, the applications escalate the process (via e-mail, SMS/app push message, console, Tivoli or log file) to the employee in charge. This person then partially or fully reverses the relevant jobs by means of a backout procedure to reset the fields to their original values. In other cases, only individual data fields are reset. This workflow results in a rapid response and immediate error correction.
A major Canadian bank operates such a standardized system (Beta 91) for automatic verification of data and logs processed on the z/OS platform. A total of approximately 400 applications are executed on the bank’s mainframe. About one quarter of them require permanent balancing – i.e., at various stages the actual values generated in the applications need to be compared with predefined control values or historic data. This process is repeated until the target and actual values match. If a process continued with wrong figures, the erroneous/outdated data could lead to incorrect calculation results or to wrong information being provided to the customer.
This setup based on fully automatic quality assurance thus helps to reliably prevent erroneous postings or, worse yet, faulty payment transfers.