Sunday, 7 July 2013

Complexity - A Meta-KPI

Complexity - A Meta-KPI

What is a Performance Indicator or a KPI? The following definition is from the Wikipedia: "A Performance Indicatoror Key Performance Indicator (KPI) is an industry jargon term for a type of Measure of Performance KPIs are commonly used by an organization to evaluate its success or the success of a particular activity in which it is engaged. Accordingly, choosing the right KPIs is reliant upon having a good understanding of what is important to the organization. 'What is important' often depends on the department measuring the performance - the KPIs useful to a Finance Team will be quite different to the KPIs assigned to the sales force, for example. Because of the need to develop a good understanding of what is important, performance indicator selection is often closely associated with the use of various techniques to assess the present state of the business, and its key activities. These assessments often lead to the identification of potential improvements; and as a consequence, performance indicators are routinely associated with 'performance improvement' initiatives. A very common method for choosing KPIs is to apply a management framework such as the Balanced Scorecard."

Specific KPI can be defined for Marketing, IT, Sales, production, etc. The problem, in any event, appears to be a subjective one, namely that of choosing the right parameters and, most importantly, understanding the business well. In KPI definition the so-called intangibles are avoided since they cannot be measured.

Science teaches us that even though everything is relative, there do exist objective and tangible properties of systems which may be measured. When dealing with mechanical systems energy is one. Mass and temperature are other examples.

Complexity is an attribute of systems which today can be measured. Growing complexity in all spheres of social life is the biggest threat to sustainable development and to a resilient economy. It would be great to measure it and to use it as a new meta-KPI. The pop-science definition of complexity, which equates it to a "process of spontaneous self-organization at the of edge-of-chaos" (by the way, ho do you measure that?) is not of much use. We propose a definition of complexity which combines topology and entropy (now that can be measured), i.e.  C=f (T; E). Let's see a example of how this definition of complexity can be used as a new KPI.
In Supply Chain Management, for example, KPIs will detail the following processes (see the Wikipedia):
  • sales forecasts
  • inventory
  • procurement and suppliers
  • warehousing
  • transportation
  • reverse logistics
Without going into further detail, let us suppose that each of the above items is already a KPI. The idea we propose is to combine these into a single holistic KPI. In other words, all of these KPIs are computed at certain intervals and the resulting data is analyzed (this is done using OntoSpace) to produce one combined measure of global performance, as illustrated in the Complexity Map below.  The map shows also which component KPIs are related to each other and in what measure and this too is very valuable.

The big advantage of this approach is that it is unnecessary to come up with subjective weights. The information of how much a particular KPI influences the meta-index is already contained in the data. In the example above, complexity (= global KPI) is, at a particular moment equal to 4.41 and corresponds to a robust and resilient system (Robustness = 89%). The question is:  how much does each single KPI contribute to this picture? The answer lies in the so-called Complexity Profile, which provides a breakdown of the total system complexity into components. An example is shown below.

We may observe that reverse logistics contributes to just over 40% of the overall complexity while transportation close to 30%. If the analysis is performed periodically, one obtains plots like the one shown below, which reflect the evolution of the meta-KPI over time.

The peaks and troughs should be analyzed (see Complexity Maps and Complexity Profiles) with care. This is because high values of complexity generally point to:
  • Increased exposure
  • Reduced profitability
  • Inefficiency
  • An unsustainable situation

The advantage of this approach is that it overcomes the weaknesses and risks inherent in (arbitrary) functions containing subjective weights. It's all in the data.