A considerable variety of Indonesian avifauna is forced into the domestic and international pet trade, where the majority of individuals are caught in the wild. To monitor the volume and development of the trade and to evaluate the threat status of the traded species, bird market surveys are usually performed. The most commonly used monitoring technique is the “Direct Counting Method – DCM”, i.e. the counting of openly displayed individuals offered for sale. In this study, we evaluate the reliability of the outputs that DCM delivers by conducting regular long-term bird censuses at two of the main animal markets in Medan (Sumatra, Indonesia) involving 10 major local vendors specialising in the Sumatran Laughingthrush (Garrulax bicolor; SL), our target species. Both markets were visited from March to December 2015 with three different survey intervals (one, two and four visits per month). In total, according to DCM, we recorded up to 461 SL individuals offered for sale. However, a comparison of the monthly logs recorded directly by the vendors during the same period revealed that DCM only uncovered a negligible proportion of the total trade. Specifically, we detected only 4.6%, 8.1% and 16.1% of the traded SL individuals in relation to the set survey intervals. While the numbers of recorded SL individuals according to DCM and the three survey intervals were significantly interrelated, none of them correlated with the real numbers of traded birds provided by the vendors. Our results suggest that census-based market data are underestimated, and represent an unknown proportion of true trade volumes, regardless of the intensity of visits. In order to obtain reliable data and prevent the underestimation of the volume of trade, we recommend of undisclosed monitoring of markets and the engagement of trusted individuals with a past personal interest in this field or, if possible, the vendors themselves.