© 2018 Elsevier Ltd Twenty mph (32.2 km/h) speed limits across urban areas are becoming a widespread tool for public health and road danger reduction globally. Determining the effectiveness of these interventions on motorised vehicle speeds is a crucial first step in any logic model that seeks to associate 20 mph speed limits to improved health and wellbeing outcomes. However, little is known about how the introduction of 20 mph limits affects speeds. This paper presents the findings from a novel comprehensive academic evaluation of the adjusted effects of a 20 mph sign-only city-wide intervention on vehicle speeds. This is based on a natural experiment that took place in Bristol, UK. Applying a quasi-stepped wedge design, speeds of 36,973,090 single vehicles, recorded by Automatic Traffic Counts before and after the 20 mph intervention, were analysed. Generalized linear mixed models were used to control for confounding variables. Results showed an unadjusted speed reduction of 4.7 mph (7.56 km/h) and an adjusted speed reduction of 2.66 mph (4.28 km/h) over two to three years. Some variability due to time variables was detected, with speed reductions being larger during night time, at weekends, and in summer months. The roads that did not receive the 20 mph intervention also saw a small reduction in speed (0.03 mph). The findings indicate that the sign-only 20 mph intervention was successful in lowering individual vehicle speeds. Policy makers are encouraged to implement a careful monitoring of the effects of 20 mph speed limit interventions on vehicle speeds in order to enable a meaningful evaluation of potential public health benefits.