Marginal Column
Prof. Dr. Felix Wortmann

Prof. Dr. Felix Wortmann
is the scientific head of the Bosch IoT lab and assistant professor of technology management at St. Gallen University (HSG). From 2006 to 2009, he was an assistant to the executive board at SAP AG. Wortmann completed his first degree and Masters in business informatics at the University of Münster, before obtaining his doctorate at St. Gallen University in 2006. His main research interests are in the areas of big data, the Internet of Things, and business model innovation. His interdisciplinary team at the Bosch IoT lab is working on issues such as smart living & working, smart mobility, and machine learning.

Copyright Photo: ITEM-HSG St. Gallen

“Data analytics enables us to achieve hybrid value creation in industry.”

Content

November 2017

 

Prof. Dr. Felix Wortmann, scientific head of the Bosch IoT lab at St. Gallen University, explains why it is crucial to always focus on a specific application when using data analytics.

Professor Wortmann, is data analytics the key to solving all problems in production?

Wortmann: No, it’s not that simple. Production has always been driven by data. Transparency is the prerequisite for improvement. I can only improve if I know what’s happening and where I am. What we are doing is increasing the resolution - collecting more data with a greater frequency. In medicine, we saw a huge advance from X-rays to CT and MRT. It fundamentally changed medicine. The same kind of thing is now happening in production with the Internet of Things, Industry 4.0 and Smart Service trends. But we’re not starting from scratch. For example, in recent decades data warehousing has taught us how to integrate data and systematically evaluate it using business intelligence. We are now taking the logical next step in the world of big data and the Internet of Things.

What is the reward for all the effort?

We are creating added value. If we fit a product with sensors and evaluate the data generated, we can use data analytics to offer additional services. People talk about hybrid value creation based on products and services. For a long time, manufacturing companies have been trying to expand their service business. It’s nothing new. However, networking of machines is now opening up totally new opportunities to embed services deep in products and to expand existing business models. Analytics is the fundamental requirement for offering successful digital services. The next major service wave will be triggered by networking of machines and data analytics. Predictive maintenance is just the start.

Is there a perfect way for users to familiarize themselves with the issue of data analytics?

In production, we still have a way to go. Of course, established Internet companies like Google have already taken the next steps. It is vital that we convince users and decision-makers of the specific benefits. It sounds obvious, but amid all the hype it’s easy to forget. In many cases, a bottom-up method is propagated these days. We collect as much data as possible and then attempt to derive potential benefits from it. I think that’s too one-sided. A top-down method can be just as effective. In this case, we first focus on the problem and specific data is then studied to identify potential improvements.

Is there a basic requirement for both methods?

Yes, it’s called agility. Even in traditional industries, we have to learn to move away from perfection, at least for a while. We need the courage to test the market with what is known as a minimal viable product, and to explore it in practice. We only look at scaling if the quick test is successful. In the agile world where we operate, exploration is no longer a closed experiment within the company, it takes place in the real world with actual customers. We need to say goodbye to some of our customary practices and established structures and be ready to go down different paths. For example, for decades conventional IT has been trimmed to achieve perfection, security and absolute reliability. It’s unrealistic to expect a complete change of emphasis overnight. It takes time. It is crucial to turn our backs on the silo mentality and to be brave enough to set out in new directions together. We need the confidence to try things out, and occasionally to fail.

With all the accomplishments of data analytics, why do we still need practical verification, for example real crash tests in the automotive industry?

If data analytics is based on models, they have to provide an appropriate description of reality. This is more or less successful depending on the application field. The complexity of a field and previous experience are crucial factors here. For example, we have come a very long way in the automotive industry. But we still need extensive crash tests today to validate our models. Weather forecasting models only work in the short term. In the longer term, the complexity is too great. There are simply too many influencing factors that we cannot accurately represent in models. A universal system that transforms a truly Babylonian diversity of influencing factors into an understandable whole at the press of a button is a vision that has been around for decades. But even the most recent achievements in artificial intelligence are a long way from turning this vision into a reality.

What is noticeable about the examples is the question of the social relevance of data analytics.

That’s right. Every technology is initially neutral. As a result, we have to actively shape our future and steer developments onto the right tracks. This needs to involve everyone, whether they are an employee, CEO, researcher, or politician.