Farm animal welfare is of increasing interest for the public and the scientific community. Therefore valid, mainly animal-based on-farm assessment systems such as the Welfare Quality® (WQ) protocol have been developed. In the present thesis 90 beef farmers in 3 countries using alternative housing systems were interviewed. Subsequently the WQ assessment protocol for fattening cattle was implemented on 63 of the farms. The aims were to (1) investigate perceptions of farmers towards the assessment system, (2) describe the welfare state of the bulls and evaluate effects of different feedback strategies and (3) describe consistency over time of welfare measures, aggregated scores and classification. Two thirds of the farmers were motivated to join a system as WQ. They expected the detection of deficiencies, advice and non-monetary benefits from its application. 92% of the farmers agreed to change management routines and 67% would increase labour input. The farmers willingness to provide data offers possibilities to reduce assessment costs. Using the WQ system, farms were allocated to the categories ‘Acceptable and ‘Enhanced. Most important areas for improvements comprised incidence of agonistic behaviours, provision of water, ease of movement and health status. After 6 months the welfare state of farms receiving a written report or additional advice did not change significantly. However, robust conclusions on the usefulness of the advisory approach taken would require longer observation periods. 6 and 2 out of 27 WQ measures achieved satisfactory consistency regarding short-term (1 month) and longer-term intervals (6 months) between visits. Considering aggregated scores, this improved to 5 and 2 out of 12 WQ criteria and 3 and 1 of 4 WQ principles, respectively. 79% and 75% of farms maintained the overall classification category. Reasons for inconsistency are most likely short-term fluctuations of farm conditions, low prevalences of clinical measures and sample size issues. For certification, frequent assessments and use of rolling averages to reliably picture welfare states are suggested. The farmers interest in assessment systems and welfare improvement coincides with deficits identified regarding potentials for WQ® implementation. Trust in the system may increase by further long-term studies on reliability aspects.