Regular welfare monitoring throughout rearing of pullets may help to identify problems early and take counteractions timely, which helps in guaranteeing good welfare. The aims of our observational study were (i) to establish and test a welfare monitoring system that can be used during (short) routine veterinary and technical staff visits for pullet flocks, (ii) to use the monitoring system to investigate variability between flocks and (iii) to analyse factors that potentially affect pullets' body weight, uniformity in body weight and mortality. The developed monitoring system tries to minimise the time required while not losing important information. Age-specific recording sheets comprise animal-based indicators of welfare and relevant environmental factors (housing, management, care) to allow for identifying causes of problems and targeted action. Finally, the system was implemented in a cross-sectional study and data collected in 100 flocks (67 organic, 33 conventional) on 28 rearing farms in Austria. Linear mixed models were used to identify factors influencing body weight, uniformity and mortality, both including all flocks (A) and only organic flocks (O) and a linear regression model with all flocks to investigate associations within animal-based indicators. High variability was found between flocks in animal-based indicators. Body weight was higher when the pre-rearing period was shorter (p≤0.001, A&O), with higher intensities of light (p=0.012, O), with only one compared to more stockpersons (p≤0.007, A&O), with a higher number of flock visits per day (p≤0.018, A&O), and a lower avoidance distance (p=0.034, A). Body weight uniformity increased, with age and decreased with the duration of the light period (p=0.046, A), and, amongst others, was higher on organic farms (farming type; p=0.041). The latter may reflect a more uniform level of welfare due to a lower stocking density and lowered effects of social competition. Within organic flocks mortality was lower if pullets had access to a covered veranda (p=0.025) resulting in an overall lower stocking density inside the barn, while in the model including all farms mortality was higher in cases where a disease had been diagnosed. We conclude that our monitoring system can easily be implemented in regular veterinary and technical staff visits, but could also be used by the farmers'. Several easy-to-record animal-based indicators of animal welfare could be analysed more frequently to increase early detection of problems. Implementation of such a routine-based monitoring system with easy-to-assess animal-based parameters and input measures can contribute to better animal health and welfare in pullets.