Studies reporting on how to control Salmonella in cattle herds have mainly been theoretical simulation models or case reports describing control of clinical salmonellosis outbreaks. The objective of this observational study was to investigate which management routines were associated with successful control of Salmonella Dublin in calves in dairy herds with previous signs of endemic infection. A total of 86 bulk-tank milk Salmonella Dublin antibody-positive bovine dairy herds were enrolled in the study in September 2008 and were all encouraged to control spread of the infection. One year later it was assessed if they were successful. The criterion for successful control was defined as the 10 youngest calves above three months of age testing Salmonella Dublin antibody-negative, indicating that exposure to Salmonella of these calves from birth until close to the day of testing had been successfully prevented. Management routines were registered through telephone interviews based on a questionnaire resulting in 45 variables for analysis. By the end of the study, a total of 84 herds had completed the interviews and had serum samples collected from calves. Data were analysed using two statistical methods: logistic regression analysis and discriminant analysis. Both analyses identified that increased probability of successful control was strongly associated with avoiding purchase of cattle from test-positive herds. Additionally, ensuring good calving area management, separating calf pens by solid walls rather than bars and not introducing biosecurity routines between the barn sections (e.g. boot wash, change of clothing) were associated with increased probability of successful control in the logistic analysis. The latter may seem illogical, but may be explained by successful herds already having good hygienic routines in place and therefore not having introduced new routines between barn sections in the study period. The discriminant analysis furthermore identified successful control to be associated with preventing cows from calving before being moved to the designated calving pen, by only letting one person be responsible for colostrum management and by not feeding poorer quality colostrum to bull calves than to heifer calves. The results are useful for dairy cattle producers and veterinary authorities to substantiate advice on management practices that are likely to lead to successful control of Salmonella Dublin.