Non-Linearity - Why should BI Practitioners know about it?
Because Non-linear nature of business is the root cause for the gap between analytics delivered by IT and positive impact of such analytics on business decisions. Let me try to substantiate that statement in this blog post.
Now, what is a non-linear system? – A non-linear system is one where the whole is not equal to sum of its parts. Let us take the example of Friction. Without friction a simple linear equation expresses the amount of energy you need to accelerate, say, a football along the ground (sounds contrived, well, it is FIFA 2010 time – Waka Waka). With friction, the relationship gets complicated, because the amount of energy changes depending on how fast the football is already moving. One cannot assign a constant importance to friction, because its magnitude depends on speed. Speed, in turn, depends on friction.
The whole body of System Dynamics developed by Professor Jay W. Forrester deals with complexity around non-linear systems. The System Dynamics Society says this and I quote – “System dynamics is a methodology for studying and managing complex feedback systems, such as one finds in business and other social systems. While the word system has been applied to all sorts of situations, feedback is the differentiating descriptor here. Feedback refers to the situation of X affecting Y and Y in turn affecting X perhaps through a chain of causes and effects. One cannot study the link between X and Y and, independently, the link between Y and X and predict how the system will behave. Only the study of the whole system as a feedback system will lead to correct results”.
As BI practitioners, we are comfortable with questions around Data Management, Reporting, Dashboarding etc. but are stumped when confronted with the question of “How do the Reports, Dashboards or any analytical artifact affect the quality of business decisions”? In my mind, the missing link is the lack of understanding of business as a non-linear system. Let me provide a concrete example here.
In one of my BI consulting engagements for a large voice-based Business Process Outsourcing company, the problem was to predict the number of calls that would be received in a particular day as that number drives a lot of decisioning on the ground, viz. number of associates required, type of skills required, infrastructure, schedule for pickup and drop etc. If this problem is taken only as a predictive analytics problem, ignoring the non-linear nature of the business, the predicted value of the number of calls (using various statistical techniques and BI tools – in this case I used the Microsoft SSAS based Data Mining solution), does not provide a complete solution. When I fed the predicted number of calls to a business process simulation model (I used Powersim in this case) that captured the inter-relationships between various business processes, a much more robust solution was obtained.
The gist of what I am trying to convey is:
1) BI Practitioners would do well to understand what non-linearity is and how businesses processes are inherently non-linear in nature.
2) Goal of analytics is to really improve the quality of business decisioning. BI does not stop with just reports, cubes & dashboards.
3) Analytics in combination with Business Process simulation models (that captures the non-linear nature of business processes) can help organizations increase the quality of business decisioning.
Thanks for reading. Please do provide your feedback.