Something new is happening in the way businesses are using information technology. After fifty or so years of digitizing every conceivable transaction, we are now on the brink of digitizing business decisions.
It is worth reflecting on the Herculean effort that has been made to digitize almost every item of transactional data. This has even extended to digitizing paper documents, nearly always carrying some form of transactional information. We developed database systems, programming languages, user interfaces, operating systems, networks, methodologies and working practices so the transaction could be digitized – with the unquestioned aim of improving efficiency. This reached a crescendo in the last decade with ERP, CRM, SCM and a multitude of other transaction oriented systems. But now the big returns are largely behind us, and we have to find another way compete.
Along with the transaction oriented systems we created an awful lot of ‘exhaust’ – in the form of data. They were mainly used for reporting – sales reports, production reports, HR reports – and so on. Always focused on the rear view mirror they gave us a view of what had happened, but no real insight into current activities, and certainly not future ones. This decade however has seen the rise of a broad swathe of technologies variously known as data mining, machine learning, predictive analytics, prescriptive analytics, and rules based systems. All these developments, and even business intelligence, have one aim – the digitization of decisions.
Examples abound – credit approval, fraud detection, sales targeting, customer recommendations, marketing optimization – and many others. They all have one thing in common – decisions are being increasingly automated, and this time it is not just efficiency we want to improve, it is also effectiveness. We want more accurate and effective decisions – greater efficiency alone is not so attractive. When we suggest additional products to a customer we want the suggestions to be relevant and attractive. It is this new dimension of efficacy that makes decision automation so attractive. Efficiency alone is a necessary but not sufficient condition for success.
What is happening right now is very reminiscent of the early days of transaction automation – fragmented systems that will ultimately become too complex to manage, and too disjoint to speak with each other. To this end efforts are being made under the banner of Enterprise Decision Management, to establish standards, practices and technologies that will help us avoid the pit of increasing entropy that affected fragmented transaction based systems.
The digitization of decisions will dwarf what has gone before. It is a quantum leap in sophistication and will create a new breed of winners and losers. Being more effective is way more powerful than being more efficient – and it will show.