Today, different analytical methods are used by different laboratories to quantify androstenone in fat tissue. This study shows the comparison of methods used routinely in different laboratories for androstenone quantification: Time-resolved fluoroimmunoassay in Norwegian School of Veterinary Science (NSVS; Norway), gas chromatography coupled to mass spectrometry in Co-operative Central Laboratory (CCL; The Netherlands) and in Institut de Recerca i Tecnologia Agroalimentà ries (IRTA; Spain), and high-pressure liquid chromatography in Agroscope Liebefeld-Posieux Research Station (ALP; Switzerland). In a first trial, a set of adipose tissue (AT) samples from 53 entire males was sent to CCL, IRTA and NSVS for determination of androstenone concentration. The average androstenone concentration (s.d.) was 2.47 (2.10) mg/g at NSVS, 1.31 (0.98) mg/g at CCL and 0.62 (0.52) mg/g at IRTA. Despite the large differences in absolute values, inter-laboratory correlations were high, ranging from 0.82 to 0.92. A closer look showed differences in the preparation step. Indeed, different matrices were used for the analysis: pure fat at NSVS, melted fat at CCL and AT at IRTA. A second trial was organised in order to circumvent the differences in sample preparation. Back fat samples from 10 entire males were lyophilised at the ALP labortary in Switzerland and were sent to the other laboratories for androstenone concentration measurement. The average concentration (s.d.) of androstenone in the freeze-dried AT samples was 0.87 (0.52), 1.03 (0.55), 0.84 (0.46) and 0.99 (0.67) mg/g at NSVS, CCL, IRTA and ALP, respectively, and the pairwise correlations between laboratories ranged from 0.92 to 0.97. Thus, this study shows the influence of the different sample preparation protocols, leading to major differences in the results, although still allowing high inter-laboratory correlations. The results further highlight the need for method standardisation and inter-laboratory ring tests for the determination of androstenone. This standardisation is especially relevant when deriving thresholds of consumer acceptance, whereas the ranking of animals for breeding purposes will be less affected due to the high correlations between methods.