- February 2, 2009
- 0 Comments
Not precisely a BPM reference, but on the Aris BPM Blog, Martin Kling posted an interesting article on Risk Management. It points out the over-reliance on a poorly understood method can really hurt your business – in this case the reliance was on VaR (Value at Risk) – which was even used to make incentive decisions for traders. As Martin points out, this incentivized traders to take positions that had a high likelihood of small gains or losses, but potentially a very small likelihood of huge losses (or gains). The small probability portion is ignored by VaR – which focuses on the 99% cases, not on the 1% cases… There’s also a link to the NYT article that goes into full depth. Both Martin’s blog and the NYT article are great reads. The tie-in to BPM? Visibility matters… If your business is relying on processes that are “black boxes” to perform well… it may be surprised when a “Black Swan” event occurs (one of those 1% events), and how much damage it can do to your business. Usually people associate “black box” with a computer program – but in this case, a completely people-driven manual process is a black box – the time delay for measurement and information to reach executives, combined with the lack of quality data, yields bad inputs to executive decision processes. BPM and other technologies can help in a few respects:
- Document the process in a non-ambiguous, but human-readable way (BPMN being the chief technology for that). This removes a bit of the black box element of process all by itself.
- Measure the process – not just the run-times of various components, but the snapshots of inputs and outputs.
- The measurement of process allows for better analysis – correlation analysis in particular – to find out what inputs drive better business.
At any rate, technological solutions should be viewed as tools to aide decision-making by participants in the process, process owners, and executives. When management starts to view the technology *as* the solution instead of an aide to capable human thought and judgment, additional risk is foisted upon the business, because software can not by nature be designed to react well to Black Swan events. But when used as a tool, software can provide the information that let’s humans sniff out the problems as they occur. (Note: in the article, it is suggested that Goldman Sachs did just that – they’re models weren’t producing what they expected by a small margin of error – so they got human risk experts together to do further analysis – which led them to reduce and hedge their positions. This is a classic case of human judgment triumphing over reliance on the model – the model was not yet showing a real problem, but it was showing enough variance for a human being to worry about it. )