The performance of dairy herds is affected mainly by factors related to cows' characteristics and herd management practices. However, these factors are interrelated, and as such, the estimation of their individual effect on the performance of dairy herds remains difficult. The aim of this study was to estimate the weight of these factors as well the interactions between them on the reproductive and economic performance of dairy farms. A stochastic dynamic model was used to simulate most physiological and management processes occurring on a dairy farm. A herd of 60 Holstein cows, with a milk yield of 8,000 L/cow-year, representative of French Holstein dairy herds, was simulated. A total of 216 scenarios were run by combining 2 levels of postpartum cyclicity resumption (average: 45 d, high: 75 d), 3 levels of 21-d conception rate of the herd (i.e., proportion of cows pregnant 21 d after insemination; low: 25%, average: 45%, high: 70%), 3 levels of probability of pregnancy loss until 120 d (low: 3%, average: 15%, high: 43%), 3 levels of sensitivity of estrus detection by the farmer (low: 20%, average: 50%, high: 90%), 2 alternative managerial goals (constant number of cows or constant volume of milk sold), and 2 types of management for the sale and purchase of animals (closed or open herd). The effect of each factor was estimated by sensitivity analysis. The parameter that had the greatest effect on reproductive performance was the sensitivity of estrus detection: a 10 percentage point increase between the low and average levels and between the average and high levels reduced the calving interval by 16 and 5.7 d, respectively. However, the factor that had the greatest effect on economic performance was the 21-d conception rate: a 10 percentage point increase between the low and average levels and between the average and high levels increased the gross margin by €62.2 and €22.3/cow-year, respectively. The pregnancy loss until 120 d had an effect on economic performance: an increase of 1 percentage point of this parameter decreased the gross margin by €2/cow-year. The other factors studied, and their interactions, did not have a major effect (low value of sensitivity indices). Closed herds or farms with a constant number of cows had economic losses of €58/cow-year compared with open herds or to farms with constant volume of milk sold. Altogether, our data suggest that, in a typical French dairy farm, farmers' efforts on estrus detection will be more profitable when associated with improvement of the conception rate of the cows.