Marketing: balancing data analytics and privacy

Data is the new oil fueling marketing activities. It helps to target the right customers at the right time with the right instrument. But recent privacy regulations limit its use. So, how do marketers make the most out of these data without breaking privacy rules? Jaap Wieringa, professor at the Faculty of Business and Economics, and his team give answer to this question by developing new marketing models that prove: targeted marketing is possible, even in the face of privacy concerns.

More and more privacy laws come into action to regulate the use of personally identifiable data. Marketers often perceive a trade-off between compliance with privacy regulations and advancing in marketing analytics. ‘Therefore, it is important to develop effective marketing techniques while staying in the realms of compliance’, says Wieringa. The research projects of him and his team focus on this balancing act between compliance and marketing, set up by two main principles: data minimization and data anonymization.

Data minimization

Data minimization refers to the principle that organizations can only use as much data as is required for the given task and that it cannot be repurposed without further consent of the customer. Wieringa: ‘You make sure that you get the most out of the data within the agreements you made with your customers.’ Wieringa’s team used data minimization for example in their collaboration with a health insurance company. Together they developed a method that can predict the churn rate: the rate of customers cancelling their contract. The new model is based on state space models, which are originally used to predict a satellite’s position. ‘The model is fed with data and after it is adjusted, the data itself is no longer needed, and can be deleted. Compared to other methods that use all available data, the privacy-friendliness of our new method does not lead to a reduction in performance, in some aspects the model even proves to be superior. For example, we can predict not only the next period, but also further into the future more accurately.’

Data anonymization

The second principle to ensure privacy is data anonymization. Especially in sensitive domains like health, it is not allowed to use personally identifiable information for marketing reasons. Therefore, Wieringa uses artificial customer data instead. ‘Analysis of artificial data can generate valuable customer insights that marketers can use to base their strategies upon.’ To generate the artificial customer data, Wieringa uses two AI (artificial intelligence) systems that compete with each other. One is simulating data with the same properties as the real customer data. The other one sees the artificial data and the real customer data and tries to identify which data points are ‘fake’ and which ones are real. The competition between the systems improves the performance of both of them. Eventually, analyzing these artificial data gives the same insights as analyzing data of real customers. What if the data is being leaked? Not a problem - the data points do not refer to individuals, because they are artificially generated.

The common ground for marketing and compliance

Privacy regulations can cause a tension between marketing and compliance departments in companies. The marketers are interested in applying new tools and methods that allow them to improve their understanding of the needs and wants of their customers. However, these methods can only be actively utilized if they fully comply with privacy regulations. At the Customer Insights Center (CIC) of the Faculty of Business and Economics, more than thirty companies work together with the researchers to find solutions to those and other marketing-related issues. Wieringa, who is also the research director of CIC, and his colleagues work to bring both parties, marketing and compliance, closer together. Their research explores new areas that excite marketers, and protect customers’ privacy. Wieringa: ‘All of our member companies are eager for this, because they are all strict in complying with the privacy regulations, but rely on effective data analysis for obtaining insights that help them to serve their customers better.’ In the end, the new marketing methods bring enthusiasm to marketing and compliance departments alike.

More information

Jaap Wieringa