During a meeting with the CIO of a Top 5 bank in Germany, I was introduced to a lady who was retiring that same weekend. The CIO averred that she was the last employee in the bank's IT department who knew the nitty-gritty of a certain mainframe application.
I pitched for a project to transform the legacy application to open system. The CIO promised to internally discuss the risks of moving from legacy to open systems and get back to me. This was in 2002 or so.
Fast forward to 2013: The bank still uses the same old mainframe application.
It's easy to conclude from this incident that banks are stuck in the old world of legacy technology. Like many others, it might even be fashionable for me to make doomsday predictions about the future of banks.
However, that would be turning a blind eye to how technology vendors have thwarted the change to the status quo.
There are at least six issues with open systems that hamper legacy transformation.
Three of them are:
- Impractical open standards. The sheer heterogeneity of banking landscapes - at least in the developed nations - precludes compliance with BIAN and other open system standards. For example, the payments landscape at a Top 5 European bank I know of
comprises 84 disparate systems, out of which not more than five can step up to open system compliance.
- Buggy code due to shorter release cycles. To quote a personal example, the social media archiving software I use keeps releasing new updates every 15 days. Most updates fail to install on the first time. This causes loss of productivity and, if the
same were to happen with banking software, tremendous reputational risks and penalties. The leading retail industry analyst RSR Research puts it nicely in its article RSR
Software Quality: What the Heck Has Happened?: "...as we move into the Cloud, I really hope we keep our feet on the ground and our heads OUT of the clouds."
- Low uptime. Not to single out Twitter, but its "fail whale" epitomizes frequent outages of cloud software. A C-level executive at a leading payments processor told me that, with the best of intentions, they weren't able to deliver more than 80% uptime.
As one who is wont to whipping out his calculator and doing quick-and-dirty calculations, I was dumbfounded when I heard this figure. Even with a 98% uptime, I'd estimated in my blog post Skating Away With Online Payments that one in 12 payments failed.
Now, if one of these systems only has an uptime of 80%, the uptime falls precipitiously to 74% (0.98*0.98*0.98*0.98*0.80) and the failure rate quadruples to four in 12 payments. In other words, one in three payments fails, which is an unacceptably high failure
rate for a mission-critical business process.
In Part 2 of this post, I'll cover the remaining three factors that hold banks back from abandoning their legacy applications. Watch this space and, meanwhile, feel free to share your thoughts in the comments below.