Digestibility and nitrogen (N) balance (NB) are classically measured in animals held in digestibility stalls for several consecutive days. These measurements are essential references for the development of feed evaluation and animal phenotyping, but raise animal welfare issues. This study quantified how the duration of the measurement period affects the precision - assessed via residual error, animal variance, and intra-class correlation (ICC) - of key traits: voluntary DM intake (VDMI), DM excreted in faeces (FDM), DM digestibility (DMd), N intake (NI), N excreted in urine (UN) and in faeces (FN), NB and finally N use efficiency (NUE). Sixteen Charolais bulls previously adapted to their respective experimental diets were submitted to two 15-day experimental periods (repetitions 1 and 2 separated by 63days) in digestibility stalls with 5days of adaptation and 10days of measurement. Eight bulls received a high crude protein diet (HIGH; 173g/kg DM), and the other eight a low one (LOW; 116g/kg DM). From d1 to d10 of measurements, VDMI, total faeces and urine excretion were measured daily. Pooled samples representative of d1-d3, d4-d5, d6-d7, d8, d9 and d1-d10 of offered feeds, individual refusals, urine and faeces were analysed for N content to allow calculation of NB and NUE for several period lengths (from 3 to 10days). Increasing the number of collection days reduced residual error for all traits, especially for NB and NUE (e.g., NB residual error for the HIGH diet: 10.8 vs 5.1g/d between 3 and 10days; NUE residual error: 4.19 vs 1.85%). Measured variables (VDMI, NI, UN, FN) were more precise and repeatable (residual error% 6-9%, ICC 38-64%) than calculated ones like NB and NUE (residual error% 15-18%, ICC 12-25%). For the LOW diet, NUE ICC increased markedly with time (from 24 to 66%), while for the HIGH diet, ICC for NB and NUE stabilised after 8days. These results indicate that at least 8days are needed to ensure reliable NB and NUE determination, especially in protein-limited contexts, and provide support for the refining of protocols to improve both data quality and animal welfare.