What the “rise of algorithmic” means for various industries is a frequently debated topic at Corvil. Over the years, we’ve worked besides many financial trading firms as they were starting to implement their trading strategies in software, meaning they were transitioning to an algorithmic way of conducting their business.
For some firms the shift was all about speed, as the simple fact is machines react to events faster than a person ever could. What we learned from them was that while speed was great, visibility got harder. In other words, things were happening so fast that it got harder to understand what was happening and why, especially when the results or outcomes (good or bad) were unexpected.
For other firms the shift was about being smarter. They had complex decision trees to determine favorable trading conditions. What we learned from them was that infrastructure performance issues impacted the trading outcomes in subtle ways. In other words, infrastructure issues would cause the algorithms to break down in unexpected ways.
So how would these effects translate into other industries transitioning to algorithmic ways of conducting their business? It’s not immediately obvious, hence our internal Corvil debates.
But one thing we know for sure: plan on cranking up your ability to monitor your algorithmic business processes. That means infrastructure, network, communications with its input data sources, communications with its output systems executing the decisions made, and anything else those processes touch.
Why? Because algorithmic business processes execute and communicate at speeds we humans are barely capable of comprehending. This means that the conditions we think are true may not be the actual conditions our algorithms are reacting to. Making those actual conditions transparent and visible to folks steering our companies is the hidden “gotcha” behind any business innovation I read about with the buzzword “algorithmic.”