or: Behavioral modeling within a bounded business context.
We humans are amazing. Truly a wonderfull mechanism of bio-enginuïty to deliver a living body that, with some basic nurtering, lives about 75 year and is able to provide all kinds of output: (not limited list:) labor, intelligence, creativity, love, hate, anger.
An important dependency for evolution is to ‘formalize’ some experienced knowledge. Most obvious example is ‘language’: the existing of language makes it possible to evolve much faster in a human life where language exists than in a ancient life where it doesn’t. The existence of the internet changed the access to knowledge in general and therefore shortens the time needed to consume that knowledge.
How wonderfull we find ourself and our evolution, it stroke me recently how strongly we evolved in some area’s (science, mathematics, bio-engineering, …) and how “un-evolved” we are at other area’s (again: not limited list) : emotions, politics, worldpeace, long-life marriage…
Some major properties we can assign to the ‘hard’-science world (and not present in the ‘soft’ area):
- unbiased data history
- symantic clarity
- general excepted formal models
- exempted from reality
The last property ‘exempted from reality’ seems obvious. We do not define mathemathical models in a holistic approach. We create a seperate world with rules and language not interrupted with the chaos-input of the world. (dixit Nick Cave : ‘Outside my window, the world is gone to war’)
Perhaps that’s why we design Business Processes as if their are extracted from the real daily business: let’s take an example:
Simple 2 step approval process.
It seems obvious because we can’t model the real behavior of this approval step, we don’t know who is, what are, how fast, …. And even if we tried, chaos/reality is changing every split second.
However, why is this? Don’t we have those properties that seem to be missing?
- unbiased data history: we are overwhelmed with data, certainly we can detect the approval point in our ERP system
- symantic clarity: unknown input can be expressed in clear averages, and be replaced in production by real time event streams.
- formal models: see below : SD is based on pure functional math and has been around since 1950’s.
The traditional BPMN model is not answering real Business Questions like :
- how many approval requests can my business handle?
- is there suffient staff for approval step 1, and if not, what is required for my business?
- what happens to my error rate if we would kill the second approval step?
We don’t talk about these questions or we try to avoid them. We leave those answers to Operational Management that simple needs to see what happens and react to it (=the behavior). That’s what they’re paid for…
The answer is to be found in the world of Systems Thinking, espacially System Dynamics, this model expresses the behavior of the process:
Run this model with a given set of values of the ‘outside’ world, the answer is defined:
Conclusion: the mystery of long life marriage is not solved, but we can model the behavior of business processes, and replace process-implementation-trial-and-error approach so we can talk about the business-stuff we can’t handle !
Spread the word.