on semantics and data intelligence


(Published in The Produktkulturmagazin issue 2 2017)

Data silos decelerate digitalisation. Content becomes isolated, they create redundancy and complicate collaboration. Semantic networks are the solution that enables companies to link information, reduce administration, and thus create time to take on the truly important tasks. Still, the times in which one finds five versions of the same file nested in complex folder structures are far from over. Which challenges will companies face, who will maintain such order during the process of digitisation?

The biggest dilemma is to be more concerned with managing data, rather than actually acting. Then there is little time left to deal with the real tasks at hand, like creating good content, for example. On top of that, parts of content will either not be up-to-date, or wrong versions may land in the wrong channels. The workload is extremely high, because one has to duplicate, transfer and correct so many things. Such retention of redundant data increases the error rate, inefficiency, and costs.

Collaborative tools are to help by improving productivity and collaboration. To what extent are these tomorrow’s data silos?

That would be the case if a tool is used to create and match data but fails to work across the entire channel as it has its own specific applications in individual channels. For example, if you use one tool to create and match online content and another tool for print content, you could bring about the exact same problem we talked about regarding file systems. Data components redundantly exist in several systems and one has to keep up with administration. Collaborative tools are good and important, but they must be used as efficiently and singularly as possible, or be integrated with the corresponding interfaces in order for data to be used across the whole system.

Your company develops software modules based on customer requirements. How can such an implementation establish efficient data handling?

The dissolution of folder structures requires a great deal of rethinking by many people as they have predominantly worked in hierarchies over the past few years. They know that you start from the top and click your way down to the target file through each folder. Nevertheless, business processes have grown over the past decades and have not suddenly turned bad just because other possibilities exist. We try to take these business processes and wrap them into new technologies and use them as best practices directly. The results that we achieve in one industry we further employ and develop with each customer. Therefore, our software modules always adapt to new requirements and are available to a broad audience.

Big Data, Analytics and Data Mining remain important tasks, which many companies do not implement, or only do so partially. Which approaches could generate new information?

New information can only be generated naturally. It is always dangerous to collect data with procedures not based on actual business processes. It is important to use a system that adapts to the company’s data situation and reflects it flexibly. This works best if a semantic network is used as basic infrastructure.

A semantic network stands for a network of terms and their relationships to each other. Could you call this a knowledge network?

Absolutely. However, we speak more of information units rather than terms. These networks should refer to business processes and represent these comprehensibly. For example, a relationship arises between a customer and a product as soon as the customer buys the product. Both the customer and the product each form an information unit that is linked through the purchase relationship. If this customer then submits a review, the semantic network extends to include the information unit ‘review’. All content can be linked via these context nodes and is therefore much easier and more intuitive as well. In addition, you avoid redundancies that usually arise within single data silos.

Semantic networks feed on data, but also on processes and architectures. What does your company do in this area?

We advise companies in the areas of process analysis, organisation and reorganisation, and on the restructuring of existing data and relationships. During project implementation, we see ourselves as architects and partners for implementation.

Many companies seek to gain information but drown in an overflow of information at the same time. How do you deal with this efficiently?

In my opinion, there is no such thing as too much information. The art is to not create redundancies and to define relevance within a question, because data that is irrelevant at one time can be found relevant in another context. It is therefore important that information can be found intuitively and easily as well.

How does semantics help to optimise cross-channel communication?

Cross-channel communication virtually requires semantic structures to be meaningful and efficient in our opinion. This means a contribution is not created for a particular channel, but rather as an information unit in the context of the whole. Thus, the same information units can be used both in a channel-crossing and in a channel-specific manner. A product text, for example, is then no longer duplicated and can still be maintained and customised for each channel individually.

Censhare CEO Dieter Reichert once called the software of the same name the “Google for enterprises”. Where are the greatest parallels?

Google has tried to gather as much information as possible and to put it in context for years. The latter is not necessarily coined to those that ask, but instead it concerns the question. For Censhare, it is similar. Asking for a product, for example, does not necessarily mean you want to know about its nutrients. Maybe you would rather learn more about the product's target group. Due to its flexible semantic data structure, Censhare helps to answer any question by showing all the content and their relationship to each other in the search.

The above software is used as DAM, PIM, MAM, MRM, CM, Web2Print, translation and collaboration tools. What is the role of semantics?

The semantic data structure has been the basis from the beginning. Without this information structure, the applications mentioned above would not be possible. Each module is based on this approach. However, the strength of the system does not come into play until the individual modules are connected. Thus, technical features from the PIM, for example, can be integrated into text for an article within content management and can be kept up-to-date.

What is your advice for companies who want to optimise their data handling?

I advise them to proceed consistently and to act non-bureaucratically in order to get rid of contaminated sites as quickly as possible. Everything is better than a classically distributed file system, in my opinion.

Loved this article? Put it in your cart!


Founded in 2004 and based in Hamburg and New York, MSP AG accompanies its customers on the road to greater efficiency and flexibility in all areas of marketing and business communication, as well as in the IT sector.


Picture credit © MSP AG

Leave a comment

Please note, comments must be approved before they are published