Predictable Business, are these only new buzz-words? (Part II)

It was very sad information. Michael Jackson is dead. But I’m not going to write the next tribute to the King of Pop here. This incident is an extraordinary example of the distinction between predictable business and simple complex event processing (CEP) or even business activity monitoring (BAM).

Simple example. Have you noticed the information that due to enormous increase of the internet traffic many systems like Google America thought they have been under coordinated DoS attack? The reaction of those systems was pretty simple: block users, slow down the answer times, use captcha test to distinguish between humans and bots and so on. It’s quite simple to imagine that for users those actions were very unpleasant and have caused deep customer dissatisfaction. It is difficult directly to calculate business loss caused by these activities but I can suppose, it is enormous. All that, due to the fact that those activities were based on simple (complex ;) event processing, simple rules that have stated “if traffic increases than block users etc.”. What a bummer!…

The idea of predictable business is much more sophisticated but the basics are pretty simple:

Predictable business concentrates on the uncertainty of the process. Only the growth of uncertainty can cause the process to be difficult to predict or even unpredictable at the end.

Let be back to Michael Jackson. Was his death unpredictable? No. Was his influence on the information world unpredictable? No. So the “unexpected” growth of the internet traffic due his death was a normal and quite obvious behavior of the process. Nothing unexpected or uncertain was there. But the reaction of the security systems was in this context completely wrong. Be careful using CEP or BAM as a base for your business decisions. It could cost you a lot of money.

The predictable business is the cure. Stay tuned…

Predictable Business, are these only new buzz words? (Part I)

No, when I see the results of global financial crisis started 2008 by bankruptcy of Lehman Brothers. Analyzing the current problems of big German companies (BMW, Comerzbank etc.) and the comments of their managers, I’m surprised how blind those managers were at the time they should have reacted. “Blind” in case of information about current situation of their companies. All manager decisions were made based on the forecast, not even on real time business data. And all of those managers were at the end 2008 totally surprised seeing the final financial reposts...
Why? How could they believe that the decisions made based on forecasts could have any chances to be right? Stupid. Even real time data is not enough to be able to plan and to react properly.

We have to predict, we need TO BE the predictable business.

And I’ll show you how to do it, to avoid such mistakes in the future.
The bases are business entropy and AI…

The first law of Predictable Business

Before I start to build the model of predictable business, I have to clearly define the term of predictability. What does it mean, that I can predict a certain event? Taken from real life, I would say, I know it will occur in the future. So to predict is to know.

As stated in Wikipedia predictability is the degree to which a correct prediction or forecast of a system’s state can be made either qualitatively or quantitatively.

The second term that requires explanation is “business”. In my context I define business from the process point of view. So to do business means to follow a certain process definition, to be a part of certain process instance. Every process contains activities and transitions between them. In a well defined process I know from the beginning conditions which are responsible for a certain flow of the process (transition from a-activity to b-activity). Such process have no uncertainty e.g. are predictable.

The first law of predictable business:
If there is no uncertainty in a process the business predictability achieves its maximum (simply saying: business is predictable in 100%).

This law is based on entropy in information technology (also known as the Shannon entropy) which states: if probability p of a discrete random variable x (in my case, the flow of a process) is p(x) = 1 than its entropy N equals to null:


But in the real life the transitions between activities are not defined as simple rules. We have to talk about probability of a transition between activities that is less than 1.

So let’s start

So, I have finally managed to start the most advanced project I’ve ever designed. In this blog I’ll be talking and trying truly to develop a framework to support predictable business… hum, what’s this? Yes, you’re right, it’s some of a kind new buzzword… but very exciting.

To build this project I’ll do my best to avoid “reinventing the wheel”, so the main task I’ve defined for me, is to evaluate existing frameworks to fulfill requirements based on identified needs.

So let’s start.