For those with short attention spans here is the central message of this article in a few lines:
1960 to 2010 – computers used as filing cabinets and calculators (ERP, CRM etc)
2010 to 2030 – computers used to find patterns (data mining, business intelligence, statistics etc)
2030 to 2050 – computers used to automate feedback loops.
For those with slightly longer attention spans here is some fleshing out of the above scenario.
For fifty years or so the dominant use of information technology has been to increase efficiency through labor displacement – replacing people with automated processes insofar as it is possible. The computer has acted as a sophisticated filing cabinet, with the added bonus that it could perform simple accounting type calculations very quickly. Systems of this nature are a cost associated with doing business and offer no inherent value to the organization. This is not to belittle the sophistication and complexity of these systems, but ultimately they offer a necessary but hardly sufficient means to compete in an ever competitive world. I have canvassed audiences in many seminars and conferences on the dominant attitude to IT in their organizations and the answer was always the same. Most businesses simply play a game of catch up without ever considering that IT might be used for something other than support for business administration. Only 10% considered that IT might have other uses.
In typical fashion we are now being bombarded with IT buzzwords which imply something quite different from the traditional use of technology. ‘Big data’, ‘predictive analytics’, ‘data visualization’, ‘data mining’, ‘business intelligence (BI)’ and so on all imply something other than admin support. These technologies, at their heart, are concerned with finding exploitable patterns. Less so with BI, but in many ways this straddles the new order and the old, and is a flawed approach in some respects (looking at graphs, dials, charts etc tells the viewer almost nothing if the probabilities behind the data are not understood) – see Fooled by Data Visualization.
Since a pattern of behavior is a prerequisite for enhancing the efficacy of a process, many organizations are rapidly assimilating a new breed of pattern finding technologies. It should be added that the patterns are not always amenable to human understanding (neural networks for example).
These are very early days in this change of emphasis, and IBM, often uncannily accurate in its marketing messages, has been using the term ‘smart’ for some time. It’s a good word to classify the new systems that deliver some form of advantage and the old ‘dumb’ filing cabinet applications.
Whether they acknowledge it or not, businesses are experiencing a bifurcation – those who embrace the new technologies and those who can’t see what all the fuss is about. And another thing that distinguishes the new from the old is the return on the investment. It is not uncommon to see organizations claiming ROI measured in the hundreds of per cent – quite different from the ROI one might expect from a new ERP suite (which might easily have been negative in any case).
Expect a rapid build up of technologies, skills, systems and capability in this domain over the next twenty years. Traditional administrative systems provide valuable fodder (in the form of data) for these pattern finding applications and inevitably will incorporate some of the new technologies.
Pattern finding is already a differentiating factor and will become much more so. There are dangers – but that is nothing new.
Many of us tend to think that patterns are long lasting things. Often they are not, and depending on the domain might last for a few years or a few minutes (capital markets). This is all pointing in one direction – to real-time markets, and the need to respond at a speed that is beyond human abilities. Implementing a pattern in production systems will obviously have an impact (positive or negative). Measuring this impact and modifying behavior accordingly is a feedback loop problem, or to use the somewhat more sexy name, a cybernetic problem. This is where we are headed, and it may happen sooner than I have estimated above, but to my knowledge none of the large technology suppliers are talking about this yet – but they will.
This is all happening with a corresponding revolution in the technology domain. Quantum computers are now a reality (see D-Wave) and it seems that Google already has one. Their use is limited but this will change. Massive computing power with very low power consumption capable of hosting advanced machine learning, cybernetic based systems. Today it is the few elite that have access to this technology, tomorrow it will inevitably become more widely available.
Everything is becoming connected with everything else. It’s a huge matrix beyond human understanding – but not beyond machine understanding. Once upon a time the means of production determined commercial privilege, but not any more. Increasingly it will be the means of understanding and responsiveness, and with it the world will undoubtedly become an even more inequitable place.