The paper analyses factors influencing the economic success of Swiss dairy farms, measured by annual income per family work unit, using panel-data regression techniques. Based on more than 5400 observations, the analysis focusses on annual milk yield per cow as key explanatory variable, adjustable by the farm manager in the medium term. We apply a random-effects model and a quantile regression based on deciles, which allows us to study the heterogeneity of the sample in more detail. Consistently with literature, the random-effects model shows a positive contribution of milk yield: an additional ton per cow results in an increase of CHF 2660, i.e. 6% of annual income. The quantile regression reveals that the impact of milk yield differs between deciles: a high milk yield is most beneficial for the best performing farms, accounting for up to 7210 CHF per ton. Our analysis further shows the influence of milk yield on profitability to be highly heterogeneous among Swiss dairy farms, indicating the demand for business-specific consulting services and not indicating the requirement for increased milk yield at each level of economic success. Key words: dairy, milk yield, quantile regression, random-effects model, Switzerland, financial performance Acknowledgement : The authors are grateful to the FADN team of Agroscope for providing the data and to Nadja El-Benni and Swetlana Renner for comments.