Abstract Background Vascular surgical care has changed dramatically in recent years with little knowledge of the impact of system failures on patient safety. The primary aim of this multicentre observational study was to define the landscape of surgical system failures, errors and inefficiency (collectively termed failures) in aortic surgery. Secondary aims were to investigate determinants of these failures and their relationship with patient outcomes. Methods Twenty vascular teams at ten English hospitals trained in structured self-reporting of intraoperative failures (phase I). Failures occurring in open and endovascular aortic procedures were reported in phase II. Failure details (category, delay, consequence), demographic information (patient, procedure, team experience) and outcomes were reported. Results There were strong correlations between the trainer and teams for the number and type of failures recorded during 88 procedures in phase I. In 185 aortic procedures, teams reported a median of 3 (i.q.r. 2–6) failures per procedure. Most frequent failures related to equipment (unavailability, failure, configuration, desterilization). Most major failures related to communication. Fourteen failures directly harmed 12 patients. Significant predictors of an increased failure rate were: endovascular compared with open repair (incidence rate ratio (IRR) for open repair 0·71, 95 per cent c.i. 0·57 to 0·88; P =0·002), thoracic aneurysms compared with other aortic pathologies (IRR 2·07, 1·39 to 3·08; P < 0·001) and unfamiliarity with equipment (IRR 1·52, 1·20 to 1·91; P < 0·001). The major failure total was associated with reoperation (P =0·011), major complications (P = 0·029) and death (P = 0·027). Conclusion Failure in aortic procedures is frequently caused by issues with team-working and equipment, and is associated with patient harm. Multidisciplinary team training, effective use of technology and new-device accreditation may improve patient outcomes.