The objective of this study was to create operational replacement guidelines under various conditions concerning reproductive performance, supply of replacement heifers and individual milk yield. Nine culling strategies were defined by three average insemination periods and by three discrimination policies between high- and low-yield cows. The effect of the nine culling strategies was analysed with combinations of heat detection rate and time of initiation of breeding after calving under two replacement heifer purchase policies: purchase (open herd) and no purchase (closed herd). The strategies were evaluated using a stochastic simulation model that simulated production and reproductive status in herds composed of dual-purpose cattle with additional young stock. The evaluation of the strategies was made in a situation without a milk quota under typical Danish conditions in 1993. Results showed that discriminating between high- and low-yield cows improved net revenue significantly in open herds but not in closed herds. Irrespective of the purchase policy, using longer insemination periods increased net revenue significantly in herds with poor reproductive performance; whereas net revenue in herds with good reproductive performance tended to increase by using shorter insemination periods. The culling rate is a poor figure when evaluating culling strategies and culling strategies should be assessed at herd level rather than per cow.