Dynamic pricing is a common practice in e-commerce. Different price claims are set depending on personal attributes of consumers – for companies this results in greater options to maximize profit. Gender-based price discrimination occurs when men and women encounter different prices for the same or very similar products and services, or when products and services are labelled as gender specific despite being the same or very similar. To analyse the prevalence and impact of online price discrimination, though, requires an intersectional approach, which focuses on how different forms of discriminations intersect and concern a person. The reason for this is that dynamic pricing for consumers is based on complex information about individuals and their price elasticity. This information is collected in real-time regarding specific products and services.
Databased pricing techniques are in part permissible under the law – also, personalized pricing if e.g. for granted discounts. The legal restrictions relate to unfair competition and the non-discrimination principle. The results of the recent survey of the Austrian e-commerce quality mark among online distributors shows that already 84% indicate to use dynamic and technically controlled pricing mechanisms. From the perspective of consumers there is a demand for transparency on prices. Personalized pricing is legally inadmissible as soon as this form of pricing discriminates e.g. by device, location, browsing and shopping patterns and gender.
The ongoing challenge is to discover and deliver scientific evidence for personalised pricing in the Austrian e-commerce sector and to analyse possible discriminatory practices in this domain. Companies act as a black-box and it remains unclear which techniques are used to constitute prices and which person-based parameters this concerns. PRIMMING aims to find evidence by developing a framework, in which personas, their behaviour and scenarios are modelled. These are to be tested automatically in controlled measurements and the results are to be compared with a control group of real-time users. The objective is to empirically determine the forms and prevalence of dynamic pricing in Austria and to further inquire into discriminations occurring in this context such as related to gender.
The results are the development of a tool to monitor static, dynamic and personalised pricing. Based on AI and Machine Learning this shall not constitute a mere observation but allow for predictions (predictive analytics). Consumers may use this tool for price comparisons, companies may use it to optimize pricing. The study which will be based on the analysed data will inform relevant stakeholders and further result in recommendations for consumers and guidelines for e-commerce providers.
PRIMMING was funded by the FEMtech programme of the Federal Ministry for Climate Protection, Environment, Energy, Mobility, Innovation and Technology.