Background: Postpartum haemorrhage is one of the leading causes of maternal mortality in low-income countries. Improving health workers' competencies in obstetric emergencies in low-income settings, has been recognized as an important factor in preventing maternal mortality and morbidity. mHealth interventions in maternal and newborn health care has shown the potential to improve health service delivery. Strong study designs such as randomized controlled trials are missing to estimate the effectiveness of the mHealth interventions. Methods: Between August 2013 and August 2014, 70 health facilities in West Wollega Region, Ethiopia were included and randomized to intervention or control in a cluster randomized controlled trial. At intervention facilities birth attendants were provided with a smartphone with the SDA installed. Of 176 midwives and “health extension workers,” 130 completed at 12 months follow-up. At baseline and after 6- and 12-months participants were assessed. Knowledge was tested by a Key Feature Questionnaire, skills by an Objective Structured Assessment of Technical Skills in a structured role-play scenario. Results: Baseline skills scores were low and comparable with a median of 12/100 in the intervention and the control group. After 6 months skills had doubled in the intervention group (adjusted mean difference 29.6; 95% CI 24.2–35.1 compared to 1·8; 95% CI – 2.7 to 6.3 in the control group). At 12 months skills had further improved in the intervention group (adjusted mean difference 13.3; 95% CI 8.3–18.3 compared to 3.1; 95% CI – 1.0 to 7.3 in the control group). Knowledge scores also significantly improved in the intervention group compared to the control (adjusted mean difference after 12 months 8.5; 95% CI 2.0–15.0). Conclusion: The Safe Delivery App more than doubled clinical skills for managing postpartum haemorrhage among birth attendants making it an attractive tool to reduce maternal mortality. Trial registration: Clinicaltrial.gov Identifier NCT01945931. September 5, 2013.