How do you move personalised recommendations away from incentivising consumption of non-sustainable products? Recommender systems have become a lot more sophisticated since they were first introduced in the 1990s, and are now widely used to personalise items on e-commerce sites and to highlight music or video based on current and previous customer choices. The challenge is that a simple model of commercial success will only measure Return On Investment (RoI) through very accurate personalisation. A sustainable alternative will use more complex data about what items are offered and the resources used both to produce them and to deliver them to the consumer so that the measure of RoI and the model of consumer choice is linked to sustainability.
Martijn Willemsen of the Jheronimus Academy of Data Science and the Eindhoven University of Technology and Christoph Ungemach of the Technical University of Munich are working on this issue. The impact of climate change makes it very relevant.
We in Coevolve IT have been working on recommender systems since 2008 when we became involved in an international collaboration that led to successful tests of personalised video-on-demand. Today you can get that from your smart TV or streamed through your laptop or phone. Personalised recommendation remains useful in many business contexts. If you would like to learn more get in touch.
Design of a test of personalised recommendation for video-on-demand service tested by BT Vision in 2010 (several product generations later now available through BT TV).