Professor Gui Liberali’s inaugural address as Endowed Professor of Digital Marketing at Rotterdam School of Management, Erasmus University described how ‘digital marketing’ has advanced to become the overarching term for the marketing of products or services using digital technologies.

In his address, Learning with a purpose: The balancing acts of machine learning and individuals in the digital society, Prof. Liberali explained how the internet has transformed some of the most basic processes in society, such as trade, payment and communication. There is more access to products, services, and opinions than ever, but at the same time behaviour is tracked more closely than ever. For example, large online retailers offer hundreds of thousands of products and can readily observe how each customers interacts with any of them in great detail. They can also rapidly deploy individual-level, in vivo, randomised online experiments at population scale to test concepts, insights and communication approaches which can lead to better services and products. However, there are often billions of possibilities, such as products-consumer combinations for product recommendations. The scale and complexity of these experiments create amazing challenges, he said.

A balancing act for firms

Firms face balancing acts. For example, they need to constantly choose between profiting from what they already know about consumers (such as the genre of movies already watched) and learning more about the same consumers (such as by recommending a movie of an untested genre). Consumers also face their own balancing acts. People inevitably leave digital footprints, but have some discretion in terms of how much information they want to keep private. Typically, consumers who are more open to sharing their preferences are also exposed to higher risks, but at the same time they can get better access to the products and services they need.

He explained that an advance in machine learning – based on predicting the best way of doing something – that focused on making sure machines know the purpose before they start the learning process, is a specific class of learning called ‘multi-armed bandits’.

He said more research is needed on traffic acquisition to tackle two major and tricky problems:

First the black box of media buying, which is related to how ad networks operate and how they sell ad space. Opening the black box of media buying would give advertisers ‘more bang for their buck’ because they would know exactly who was watching their ads and could target at a much more individual level, giving more informative advertising.

Secondly, more research is needed on the balance between attracting consumers to the store and improving store performance, or ‘conversion’ which can be deceptively hard for practical reasons. Consumers see these as two different processes, and they present two interesting research problems; allocating budget and resources to ad campaigns that bring traffic to a store, or to the store itself, which can convert traffic into sales; and attracting traffic that would involve a design campaign problem to avoid inconsistencies between the contents of an ad and what’s found in the online store. More research is needed on understanding and reconciling need for consistency as well as for differences in what information is needed, said Liberali.

Prof. Liberali intends to tackle this problem with Prof. Gerrit van Bruggen and PhD student Zhou Xin, by expanding the scope of ‘morphing’ methods to include attraction and conversion. Morphing uses a structure of cognitive styles to identify and deliver the best version of the website to a particular user. Liberali’s research shows that it can increase sales performance by 20 per cent.

Prof. Liberali also sees further application of his morphing models in the creation of medical treatments and drugs. They allow drug companies to develop new treatments using smaller groups of patients – thereby putting fewer people at risk – while getting new treatments to the market more quickly.

Dean of RSM, Prof. Steef Van de Velde sees Prof. Liberali’s appointment as logical and necessary. “While we are a school of management at heart, our students want to understand computer coding,” he said. “Gui Liberali is the one who can take the lead in subjects like machine learning, big data, and algorithm-building, and then link that to business practice.”

Professor Gui Liberali has collaborated in research with firms in the US and Europe and has presented his research at leading research universities in the US, Europe and Australia. He was the first to have marketing students design and print 3D prototypes in classes about development and launching of new products. He currently teaches courses about innovation and new products for RSM’s MSc International Management/CEMS programme, as well as minor courses about digital marketing techniques such as text mining, A/B testing and website morphing that have attracted students from RSM, the Erasmus School of Economics, and Delft Technical University.

Digital marketing is an emerging and lively scientific field that has become focus of much attention and a major topic for managers as well as scientists. RSM’s new chair in Digital Marketing is aimed at contributing to the production, application, valorisation, and dissemination of knowledge in digital marketing. The field addresses questions and challenges that are not tackled by traditional models and theoretical frameworks. Firms and consumers now increasingly rely on the internet, mobile technologies, wearable computing, and visual and speech recognition. There is also the issue of finding the balance between bricks-and-mortar shops and online stores, or allocating marketing budget between online channels and traditional channels such as newspapers, radio and TV.

Dr Liberali’s work has appeared in a variety of journals, including Marketing ScienceManagement Science, the International Journal of Marketing Research, the Sloan Management Review, and the European Journal of Operational Research. He is also Vice-President for Membership of the INFORMS Society for Marketing Science until 2019.

See the original article at: