The idea that an Enterprise can be modelled on the human mind isn’t new. The term digital nervous system was originally used as far back as 1987 but it was made famous by Bill Gates in his 1999 book, Business @ the Speed of Thought. However even back then The Register pointed out the flaw in this approach.
In trying to apply the concept to computing, you come unstuck very quickly because you can’t validly compare a system where the outcome is determined by logic and the information content, with a system where the outcome is determined by evolution.
It is precisely this argument that has spiked my interest in the idea.
First some context. My interest here is in how the customer facing elements of an on-line enterprise can respond to the individual needs of each customer, now and well into the future. In particular how we can avoid the need for a constant cycle of too-big-to-fail re-platforming projects. Rather have a platform that can evolve. The holy grail here is a platform that not only bounces back every time there is a need to change, but rather bounces back stronger, fitter than it was before, having learnt more about the type of changes it can expect in the future.
So how do we create a learning customer experience platform? I’m going to begin with the thesis that if we are to create a platform that can learn, we should model that platform on the way we think. The psychologist Daniel Kahneman (Nobel winner in Economics) who talks about having two systems that we use for thinking. “System 1” is fast, instinctive and emotional; “System 2” is slower, more deliberative, and more logical.
System 1: Fast, automatic, frequent, emotional, stereotypic, subconscious
System 2: Slow, effortful, infrequent, logical, calculating, conscious
So what does it mean to take this model and apply these ideas to a customer experience platform?
System 1 is based on rules. Heuristics based on experience of working with real customers. These may not be the most optimal solution, but they will be fast, able to work with very limited information and critically not require the system to know the individual customer history. The response of System 1 is based on the behaviour of the customer in the here and now. Rules will evolve quickly using a basic measure of “fitness”. For an e-commerce site rules that result in larger sales will thrive, rules that have a negative impact will be culled.
System 2 is based on data. This is where analysis of the data held on the customer happens. Analysis is based on historic data. Trends and patterns emerge and are used to refine the experience. Users can be targeted on an individual basis. This is typical to the approach taken by Amazon, e.g. recommendations, related items. Classic big data.
If all your customers login before any other interaction then you can possibly get by with just System 2. However for most enterprises the trick is knowing when to use each system. There can also be a flow of rules being created in System 1 based on the analysis taking place in System 2. The other factor, is that both Systems will require human interaction to help with rule creation and data analysis, However both systems are evolutionary in nature, and while the framework can be “designed” for holding the data, and running the rules, the resulting customer experience will emerge via feedback with real customers.