The TGOA university defines a new unique knowledge platform bringing teachers and students together in order to rethink and redefine the processes of digitalisation
polyphonic data model 18
By Temel Kahyaoglu
The time has finally come. We believe that Europe has always been a master of polyphony. Not just in terms of music, but also – and particularly – in the way we understand things, structures and democratic coexistence. If we look at the B2B goods sector in Germany in isolation, it becomes clear why we as a nation are praised – but also criticised – throughout the world for complex thinking.
However, if we look at the communicated simplicity of an iPhone, the seemingly not-so-simple Tesla Model 3 production line or the inability of Facebook and Google, etc., to competently deal with anti-democratic fake news, it becomes clear that we all love the simple. We just do not like things to be complex. And should us Europeans specifically think more simply in order to be more successful?!! Which is why everybody working in communication uses this paradigm. I am also a firm believer in simplification. Marcus Aurelius, the last of the stoics, correctly stated that, when reduced to their nucleus, the most complex of elements – and, as a result, ultimately also merged elements – become understandable even to mere mortals.
If we apply this knowledge to digitalisation, it corresponds to the way we dealt with computers and applications until recently. We defined simple procedures. Usually procedures that we would also have described and implemented in this form or similar in our biological world. Then we created systems that can carry out these procedures faster and in immense quantities, reminiscent of the very first macros in Excel. We also liked this, as we were able to understand more and do so faster by means of neurological networking.
But it has been other viewpoints that have created innovation in industry, particularly in Europe. The genius of thinking in new synapses has been an invaluable character trait of the most successful men and women over the past 200 years.
Even if many media offer a different take on current innovation, I have noticed in many discussions with representatives of this industry over the past 12 to 18 months an unbridled energy and, above all, the attempt to use digital possibilities to no longer replicate, but to create new networks that require new perspectives, and hence new synapses. This is hugely exciting, and we at The Group of Analysts have been observing this new context affinity for some time now. But it was only during the analyst meetings held at the end of the last year that we developed a follow-through and the associated methods for digitalisation. We invented the term contextual intellectual property back in mid-2016 in order to direct the focus on to networking and the resulting benefits. However, it was only at the beginning of this year that it became evident that this has to be started not as late as the networking stage, but as early on as during the information creation and structuring phases. And this was the very moment that the term ‘polyphonic data model’ was born.
As in the case of a good orchestra, each individual instrument has to function independently. With specific skills, a brass instrument can be seduced into releasing fabulous sounds, but this ability is not necessarily imperative for a string instrument. Both instruments, viewed as single entities, are fascinating and are able to exist on their own. However, combining the two can result in a new fascination. If we apply this to many instruments, the outcome can be wonderful, but also unbearable. And if we apply the polyphony equation to digitalisation, it quickly becomes clear which ones we have to first create in order to be able to conduct an orchestra. The individual instruments and associated musicians all have to be individually perfect and conducted in concert. Like a conductor thinks and controls using a neurological synaptic structure unknown to mere mortals, we all must first see the new potential of the digitalisation polyphonically and allow these new synapses to first develop, before deploying them across the board.
But let us be a little more concrete. Creating information particles in digital systems always simultaneously creates the associated structures, the so-called classes. For example, we might create a product structure within a business, in other words, an entry into the business segment, product families and products – all the way through to the article and/or part number. These classes are then augmented with attributes, for instance E-Class, diesel, convertible, etc. In addition to classes and attributes, we have people. So, we define who can see and do what. This represents the classic PIM (product information management) system (hugely simplified). This product structure has been frequently referred to as a master structure or master class. And it is precisely this class that generated all possible deliveries of data. However, orchestrating a large number of channels from a master class is not what we are talking about. We simply believe in a different master. It is not the class itself that is the master, it is rather the network of numerous, independently-functioning classes that is the master. This means that we are not allowed to describe a relationship between two classes in simple terms, but must allow the relationship to be described individually and additionally also provide a view of the descriptions of the classes it is related to. This can be unlimited from the perspective of the relationship. Whenever we talk about the figure of eight, our imaginations start to go on something of a roller coaster ride. But that’s a good thing, because it is the only way we can exploit digitalisation for ourselves and achieve new views and synapses that we would be unable to master on a purely mathematical level.
But let us return to our example of the product class with the Mercedes. In a polyphonic data model, you would now create a further structure/class for car accessories and after-sales products, for instance. To this end, roof-mounted carriers and original air filters would be self-sufficiently depicted in autonomously-functioning classes. A further class might depict the customer structure. In other words, a structure of all E-Class drivers and potential customers. Here, the attributes would then be classical CRM attributes such as name, address, title, etc. If we now further extend this principle, the ideas for new classes are virtually unlimited. We could also create a further class for social media comments, i.e. all comments relating to the E-Class posted on Facebook, etc., would then be depicted in the class.
I think the principle is clear here. From the perspective of polyphony, we therefore talk about creating anything from several to an unlimited number of classes/structures. In themselves, these have their own, independent attributes – in other words, descriptions – and can be viewed, processed and made available by dedicated people. Polyphonic classes are therefore not limited to a product domain, but can be defined as desired in terms of contents. Here, the TGOA analysts go the corresponding step further than our American counterparts at Gartner. We do not believe in a master data class, which among other things represents the product class, we believe that the polyphonic class is not the focus in terms of contents. It only becomes truly interesting when the polyphonic classes are networked among each other. In other words, when the E-Class has a relationship with a potential customer and/or the roof-mounted carrier and/or air filter and/or Facebook user comments. And we are not just thinking merely of networking, we are naming the relationship as the master. Consequently, we want maximum technical potential for the relationship. In other words, attached to the relationship are independent attributes, independent associations to people who can see and process this relationship and have access to it. Plus, all attributes, class information and rules relating to people, which show from the relation to the polyphonic classes are also made available at the relation. A relationship, for example, to 4 polyphonic classes, can therefore represent 5 different attribute models and rules relating to people. A relationship with 100 polyphonic classes can represent 101 attribute models and rules, etc.
Technically, the networks created are not tangible in terms of our neurological thought patterns. For a cable manufacturer, potentially with more than 1 million items in its portfolio, expansion using polyphonic classes can create a network that might, on the one hand, seem impossible to visualise and, on the other hand, appear anything but simple in its complexity. But this is precisely the idea. We create these networks by means of the intelligent networking of polyphonic classes. This is what we mean by a polyphonic data model. We focus less on controlling an individual master class/master structure and more on developing data models that create the networks between classes. Even though this may be complicated to read in the above, when applied, polyphonic data models are relatively simple to use. For example, correlations between E-Class, specific engine, customer, lease features with regards to mileage and the relationship to the respective air filter can optimise an order for the nearest customer service site in a timely manner; the perfect solution for all parties involved. Heavily-hyped terms such as ‘predictive after-sales’ can be made usable for all us mere mortals with a simple polyphonic data model rule.
Whenever I am asked about AI – or artificial intelligence – I always respond by saying that, for me, AI stands for ‘Absence of Interest’. This is, of course, a joke, but expresses precisely the way we complicated Europeans think. Don’t just talk the talk, but master complexity using tools. We already have pattern recognition systems that can quite rightly call themselves artificial intelligence. In complex classes that – networked among each other – result in real intelligence, we are currently all at the very beginning of a new age. However, we also see some manufacturers, consultants and integrators promoting the right topics. Whether topological changes in databases to gain control of this new complexity, open entity models that enable independent and self-sufficient descriptions in polyphonic classes, or whether there is merely a lack of openness among businesses operating in collaborative networks – all these make me personally look positively towards the future, and I therefore recommend that we all interact with each other with enjoyment, inquisitiveness and drive. Speak to our analysts – we have all mastered at least one instrument. The age of silos and solos has ended.
Temel Kahyaoglu is the CEO of The Group of Analysts AG. As Chief Analyst, he studies the topic of data modelling in great depth.
This article was published in The Produktkulturmagazin, issue Q1 2018. Picture credit © furtseff, Courtesy of Shutterstock