The true incidence of bovine abortion (42–260 days of gestation) is unknown because of under‐reporting. This is particularly true of the first half of gestation when most abortions occur. From research studies, it is estimated that the animal‐level incidence of abortion between day 45 and term is approximately 10%, being lower in suckler than in dairy cows. While wide variation exists in herd‐level incidence, up to 30%, it is estimated that less than 5% of farmers report high (>5%) observed abortion rates. Given these figures, veterinary practitioners and farmers tend to use investigation thresholds close to these values, ≥5 and ≥3%, respectively. Most of the risk factors for abortion are animal‐level factors (e.g. parity), but there are also herd‐level factors (e.g. enterprise type). Of the many risk factors for bovine abortion, only a small number are modifiable. While the causes of bovine abortion are often classified as noninfectious or infectious, in reality, this means diagnosis not reached (DNR) or infectious. Approximately 40% of abortions are diagnosed as infectious and the remaining 60% as DNR; DNR includes both infectious and noninfectious causes not diagnosed. Increasingly, genetic causes are being diagnosed and changes to current national genetic selection indices may be warranted. Of the infectious causes, Neospora caninum is now the most commonly diagnosed abortifacient internationally, followed by Trueperella pyogenes, BVDv and fungi. Both national eradication programmes and effective vaccination programmes can reduce the burden of some infectious abortifacients, but the control of sporadic, noninfectious and Neospora‐associated abortions remains challenging.