There is little evidence that the productivity of New Zealand beef herds has improved over time. Data from the NZ Meat and Wool Board’s Economic Service (2006) suggest that the average national calving percentage has declined over the last two decades. During the same period cattle carcass weights have increased but so too has the average cow live-weight which has resulted in increased maintenance costs of the cow herds. It is unclear whether production efficiency in the industry has improved or declined over time. The aim of this research was to develop means of improving productivity in commercial beef herds through practical methods of performance recording. The objectives were firstly to establish current management practices in commercial herds and secondly to develop an objective system for cow selection and culling which would have practical application in commercial herds. Beef management survey Ninety two commercial beef producers with more than 100 breeding cows from the greater Canterbury region of New Zealand were surveyed. Pasture control was the main reason given for owning a beef herd. Size and conformation were the main selection criteria for choosing replacement heifers and bulls. Over 80% of herds retained their own heifers as replacements and >60% mated yearling heifers to first calve at two years of age. Fertility was poor in the surveyed herds. In-calf rates at pregnancy testing averaged 88% for maiden heifers, 92% for rising second calvers and 93% for mixed age (m.a.) cows. There was no significant difference between in-calf rates of maiden heifers mated to first calve at two or three years of age; nor was there any significant difference between the re-breeding success of the two groups. Heifers mated at least one week earlier than m.a. cows, achieved a re-breeding success 4.7% greater (P<.01) than those mated at the same time. Reasons for cows not weaning a calf included wet dry (9.3% of pregnant cows wintered), pregnancy tested not-in-calf (7.4%) and dam death (2.6%). Only 87.9% of pregnant females wintered weaned a calf (89.4% of m.a. cows and 84.9% of heifers). Reasons why cows exited the herds included diagnosed empty (37.2% of all exits), involuntary culls (25.4%), sold wet dry (16.2%), deaths (13.1%) and poor calf production (5.1%). Vaccination was infrequent with clostridial vaccines the most common in m.a. cows (15.2%) and in calves (40.7%); vaccination against Leptospirosis was much less common. Very few of the surveyed farms used any system of performance recording; as a result there was very little performance-based selection or culling practiced. Evaluation of alternative measures of cow productivity Data from four performance recording beef herds were used to compare alternative measures of cow productivity with the industry standard which is calf weaning weight adjusted for sex (SOC) and age of calf and age of dam (AOD), i.e. the “200 day weight.” None of the alternative measures evaluated required knowledge of calving date and all were relatively easily obtainable in extensively managed beef herds. The assessment of cows was based not on their estimated breeding values but instead on their most probable producing ability which, as the sum of all of the permanent, repeatable aspects of the calf-rearing ability of the cow, explains considerably more of the variance of weaning weight than does breeding value alone. SOC and AOD-adjusted marking weight, weaning weight and average daily gain (ADG) between marking and weaning were the traits mostly highly correlated with the 200d wt of calves (r = 0.68, 0.90 and 0.74. respectively). An Extensively- Grazed-Cow-Weaning-Index of these three indicator traits was found to be more highly correlated than any of the individual traits on their own (r = 0.94). Index weights for the three indicator traits were calculated within each herd and then those within-herd index weights were regressed on readily obtainable herd descriptive variables to obtain a regression equation that could predict index weights for any herd. When the model was applied to data from two additional herd years not included in the original model, the EGCW Index was highly correlated with the 200d weights (r>0.90). Performance-based culling of previously unselected commercial beef herds based on the EGCW Index will result in improved productivity due to the moderately high repeatability of calf weaning weight. Objective data from extensively grazed commercial herds will also make possible the use of commercial herd data in genetic evaluations of herd sires.