Predictive maintenance is gaining more and more importance. However, current research and implementations mostly focus on industrial applications. As high quality sensors are getting more precise and powerful, and at the same time are getting cheaper, alongside with more advanced and ubiquitous means of communication, predictive maintenance use cases in the consumer good area raise and start being more feasible. In this paper we present an idea of predictive maintenance on bike chains, applied on two typical major use cases. The proposed approach is based on solid-borne sound detection and digital analysis of chain malfunction from different causes. Finally, we identify real world use cases and appealing business models to be applied on the methodology.