부산대학교 도서관

The pay-as-you-go (PAYGo) model is now the principal way through which solar home systems (SHSs) are distributed in Sub-Saharan Africa. By alleviating the upfront cost and providing flexible payment schemes, the PAYGo model helps tackle what is still the main barrier for SHS adoption-i.e., affordability. However, the scheme’s design and evaluation are still largely guided by assumptions on user behaviour. This work provides a first evidence-based look into SHS PAYGo user payment patterns and behaviours, by using payment records of over 32,000 Rwandan SHS users. Three clustering algorithms are implemented to conduct a customer segmentation, employing an ensemble validation method which facilitates qualitative oversight. The analysis reveals five user payment behavioural profiles which serve to aid improvement in the current PAYGo model design.