The dairy sector is urged to reduce environmental impacts, such as greenhouse gas (GHG) emissions. But dairy farms not only produce milk: surplus calves and culled cows also yield meat as co-product. To split environmental impacts between milk and meat, a biophysical allocation method proposed by the International Dairy Federation (IDF) is currently used. Its applicability to farms with large meat-to-milk output ratios (beef-to-milk ratio, BMR) may be limited and lead to wrong conclusions when assessing GHG emissions and mitigation measures at farm level. To overcome these limitations, we developed a biophysical allocation a approach based on the net energy requirement for milk and meat production according to internationally agreed energy requirements for dairy cows. Both the enhanced and the existing allocation methods were tested on an international dataset that included farms with a large range of BMR, as can be found in dual-purpose production systems or on farms with low milk productivity. The results from the international dataset reveal that the allocation factor does not substantially change for production systems with low BMR. For BMR up to 0.03 kg live weight (LW)/kg of fat and protein-corrected milk (FPCM), the maximum deviation in the allocation factor between the two methods was 0.047. For larger BMR, the developed method still allocated relevant shares of emissions to meat while the standard approach did not. The developed method is less sensitive to shifts in BMR, especially for low-performing dairy farms. In addition, both methods were tested on a dataset of 46 Swiss dairy farms. By increasing the longevity of cows (one additional lactation), the impacts of altered BMR on the modelled GHG emissions and their allocation on milk and meat could be assessed. Increased longevity resulted in fewer cows to be replaced, decreased emissions from the rearing of replacement stock (-444 kg CO2-equivalents/cow/year) and lower meat output (-61 kg LW/cow/year), as fewer cows were culled. Consequently, a larger share of emissions was allocated to milk. While the standard biophysical allocation approach did not result in reduced GHG emissions per kg of milk (+0.002 kg CO2-equivalents/kg FPCM), the newly developed approach generated a modest (-0.022 kg CO2-equivalents/kg FPCM), although not significant reduction.