In 1985, external quality assurance was initiated in the Netherlands to reduce the between‐laboratory variability of leukemia/lymphoma immunophenotyping and to improve diagnostic conclusions. This program consisted of regular distributions of test samples followed by biannual plenary participant meetings in which results were presented and discussed. A scoring system was developed in which the quality of results was rated by systematically reviewing the pre‐analytical, analytical, and post‐analytical assay stages using three scores, i.e., correct (A), minor fault (B), and major fault (C). Here, we report on 90 consecutive samples distributed to 40–61 participating laboratories between 1998 and 2012. Most samples contained >20% aberrant cells, mainly selected from mature lymphoid malignancies (B or T cell) and acute leukemias (myeloid or lymphoblastic). In 2002, minimally required monoclonal antibody (mAb) panels were introduced, whilst methodological guidelines for all three assay stages were implemented. Retrospectively, we divided the study into subsequent periods of 4 (“initial”), 4 (“learning”), and 7 years (“consolidation”) to detect “learning effects.” Uni‐ and multivariate models showed that analytical performance declined since 2002, but that post‐analytical performance improved during the entire period. These results emphasized the need to improve technical aspects of the assay, and reflected improved interpretational skills of the participants. A strong effect of participant affiliation in all three assay stages was observed: laboratories in academic and large peripheral hospitals performed significantly better than those in small hospitals. © 2015 International Clinical Cytometry Society