ALGORITHMS, AI AND PRECISE METHODS: HOW MATHEMATICS CAN SUPPORT MARKETING

Scherfner's interest in mathematics was sparked by a visit to the Natural History Museum in Salzburg, where he was inspired by posters about the work of Albert Einstein. Today, he teaches mathematics and theoretical computer science and advises scientists on their career paths. In addition to his professorship, he is also an author. In this interview, he reveals how exactly mathematics and artificial intelligence can be used in marketing.

In the interview, he talks about how mathematics and artificial intelligence are changing marketing – and how people are still involved.

BN: How can math support the work of marketers?

Prof. Dr. Mike Scherfner: Mathematics is an important tool for modeling in many classical sciences, from biology to physics to sociology – so it's no surprise that it has a lot to offer modern marketing as well. Many branches of mathematics are applied here, for example statistics is used to analyze large amounts of market data. This helps to identify trends, understand customer behavior and make informed decisions. Overall, mathematics transforms marketing into a data-driven, precise science that enables strategic decision-making and significantly increases success.

Let me give you a more specific example: I would like to focus on the area of AI that is driven by algorithms and based on statistical methods. One subfield deals with structures that represent significant parts of our brain (and also that of animals) as a mathematical model in the form of a so-called neural network. This is designed in such a way that it can learn from examples – for example, to distinguish between bumblebees and cherries. Even the creators of these complex structures are often no longer able to see what exactly happens when the algorithms are working. However, we are quite good at determining the precision of the results – in other words, whether the correct answer is given when an image of a cherry is shown. But we don't have to think about the difference between cherries and bumblebees, because patterns can also be found in purchasing behavior, and that is precisely what now helps marketers make decisions.

BN: One big issue is the “black box” customer – why we buy something or how we react to advertising messages, for example, cannot be fully explained. What answers can mathematics offer for such problems?

Prof. Dr. Mike Scherfner: By analyzing large amounts of data, mathematics can identify patterns in purchasing behavior, as described above; even if the exact motivations of an individual customer remain unclear. Furthermore, AI is also partially able to recognize structures that remain hidden to us. For example, it seems to be possible for pattern recognition systems to identify the differences in terms of “female” or “male” when looking at photos of retinas. A capability that the human observer does not have, as a specialist on the subject told me. Thus, in many cases, we can now place a certain amount of trust in the results of AI, even if the path to them (see above) and the result itself are not always completely clear to us.

But of course there is more to it than that: instead of trying to fully understand individual customer decisions, the mathematics can calculate probabilities. This makes it possible to at least estimate how likely it is that a customer will respond to a particular type of advertising or purchase a product. It is often possible to identify overarching trends that provide good insights into the behavior of entire target groups. Algorithms enable marketers to make predictions about how certain customer groups will react to advertising messages. These models are based on historical data and help to better predict future reactions.

BN: Is artificial intelligence a threat or an opportunity for marketing occpuational profiles?

Prof. Dr. Mike Scherfner: Such questions are usually (in my experience so far always) answered in such a way that the human (here the marketer) remains the winner in the end, whose abilities will be needed for all eternity. However, I am very cautious about this. If you look at the history of AI, you will notice that there have been huge leaps again and again that could not have been foreseen. After all, marketing is not done for machines. From this point of view, humans will therefore be part of the whole and may continue to make decisions. But will these decisions be better for their own goals than those made by AI on the same topic? I don't think so. My guess here is that in the not too distant future, AI will replace various human workers and will be able to do things that we can't even imagine today. However, this does not yet say anything definitive about the benefits or harms of AI – we should just let time do that.

BN: How can an understanding of math enable future marketers to succeed in their careers?

Prof. Dr. Mike Scherfner: I am convinced that essential skills should be in place to understand the basics of the respective current technologies. For the methods of AI, special knowledge of stochastics (the generic term for statistics and probability theory) should be included, because this also helps to avoid thinking traps and gives you a feeling for probabilities, the importance of which has already been mentioned above. By regularly dealing with mathematics, even if it is only regularly estimating everyday calculations, you develop a kind of intuition that can also warn you of misjudgements in good time – which is not only helpful in marketing.

BN: How can you inspire enjoyment of mathematics?

Prof. Dr. Mike Scherfner: To be honest, our brain is not really keen on mathematics in terms of its development, because it was created for solving other problems. Now it could be said that in mathematics, after all, problems are also solved. Absolutely. But here it is usually really about very complex issues, the processing of which requires a lot of energy (the brain is a big fan of glucose anyway). But using up energy is not very popular, because our inner program comes from a time when energy was in short supply. And then use the little that is available to solve abstract mathematical problems? Probably not. Therefore, I suspect that a love of mathematics is simply a kind of (wonderful) defect that you can't invent for yourself. Nevertheless, it is certainly possible to approach joy, for example by being a good role model. Those who are enthusiastic can be infectious. Furthermore, finding solutions generates a little happiness (keyword: reward system). It is also good to see how you can achieve success in your own practice by using mathematics.

BN: In what area are you currently doing research at the Anhalt University of Applied Sciences?

Prof. Dr. Mike Scherfner: There are two areas that I have been in love with since my student days: geometry and relativity (to use rough terms), and these are the areas in which I continue to work. I am particularly interested in mathematical considerations that lead to so-called cosmological models in which time travel is possible. Over the years, I have increasingly dealt with didactic topics and have also recently written a paper on this subject with a colleague. The topic remains exciting! In my free time, I and some dear colleagues keep writing books on mathematics, such as a work on the mathematics in the bachelor program in three volumes (see link below). Right now, I'm working with a friend on a book about mathematical aspects of traffic. This book shows in a rather humorous way how much mathematics is actually involved in things like curves, transmissions and traffic rules, and it conveys a lot of what people want to know about mathematics when they start college.

Links:

• Marketing & Communication Management FT
• Digital Marketing FT
• Book tip: Mathematik für das erste Semester (Scherfner/Volland)