What Does Behavior-Based Architecture Really Mean?

Sometimes I see a customer doing a major IT transformation or launching a major business model transformation, but very rarely do I see a customer doing both at the same time. One mobile carrier in the U.S. is attempting both with key initiatives spanning the  consumer perspective, front office, back office, and their network, making for a project with many moving parts.  

Move From Transaction to Behavior Architecture

The image above is the whiteboard I use. It is a reusable image that shows why it is essential for transforming telcos (retailers, banks, airlines) to move from a transaction-based architecture to a behavior-based architecture. The term “always late architecture” describes the traditional transaction-based architectures.

The whiteboard has two sections, left and right. The left side shows how traditional IT infrastructures have been built over the past few decades. They are all transaction-based architectures. An event happens (my wife goes into a store), making a transaction that is stored in mainframe/ERP/DB. The action happens in the mainframe/ERP/DB. Analytics are run on the mainframes/ERP/DB and then a letter is sent via mail to my wife with an offer.

Now, almost all businesses need to transform to a behavior-based architecture. A consumer (my wife) is at the retail store with the retail app and she is asking, “Mr. Retailer, I am telling you my location and you have my profile information. Do you have an intelligent response for me?”

Take Action Right Now

The transformation from the left to right shows a major paradigm shift in deciding where to take the action. The action moves to the real-time in-memory business layer that combines data, events, and analytics.

The transaction will still eventually happen in the system of record (mainframe, ERP, DB), but most of the valuable business optimization capability moves into the TIBCO layer.

Left Side:

  • Not elastic (You cannot easily expand and shrink mainframes/databases.)
  • Not scalable (These technologies store transactional data, NOT massive amounts of real-time events.)
  • Always late (Response and analytics are not a part of the business processes.)

Right Side:

  • Elastic (Grows and shrinks based on your business needs.)
  • Scalable (Handles millions and billions of events.)
  • Real time (The architecture combines data, events, and analytics to optimize business processes continuously.)