Crash blackspots significantly impact and, to some extent, determines the entire road network’s safety level. Therefore, it is imperative to identify these blackspots and investigate the contributing factors. This becomes particularly crucial for low-income countries facing financial constraints in implementing road safety measures. Methodologically multiple studies utilised random parameter negative binomial models to predict vehicle crashes due to their ability to address unobserved heterogeneity in crash data, surpassing conventional models. However, the potential of this promising method in investigating factors influencing crash blackspots remains underutilised. This study aims to identify crash blackspots and investigates the roadway factors of such segments using the random parameters negative binomial modelling method. A three-year (2017–2019) crash data collected from the Ethiopian capital, Addis Ababa, with traffic volumes and various geometric characteristics were utilised. The model estimation results demonstrate the superiority of the random parameter negative binomial model over conventional models, showcasing its ability to reveal unobserved heterogeneity associated with road condition factors in crash blackspots. The study finds that horizontal curves and access density are significant road condition-related contributors to crash blackspots, characterised as random parameters. On the other hand, fixed-parameter influence factors include average annual daily traffic, vertical gradient, vertical curve, median width, and traffic control devices. The study highlights the need to further explore horizontal curvatures and access control as potential random parameters in crash blackspot locations. The findings may assist transportation planners/agencies in prioritising road maintenance, enhancing design standards, and implementing targeted safety interventions to improve road safety effectively.