OBJECTIVE: To determine whether elective frozen embryo transfer (eFET), or the 'freeze-all' strategy, associated with better cumulative clinical outcomes compared with fresh embryo transfer (ET). METHODS: A total of 7,236 IVF cycles that were followed by a fresh ET or eFET between 2013 and 2017. The patients were subjected to controlled ovarian stimulation (COS) with a gonadotropin-releasing hormone (GnRH) antagonist protocol and cleavage-stage ET. Embryo cryopreservation was performed on day 3 by vitrification using an open system. A comparison of cumulative outcomes between the eFET (n=4,065cycles) and the fresh ET groups (n=3,171cycles) were performed. The analysis was performed in four groups of patients based on the number of retrieved oocytes: Group 1: poor responders (1-3 oocytes); Group 2: suboptimal responders (4-9 oocytes); Group 3: normal responders (10-15 oocytes); and Group 4: hyper-responders (>15 oocytes). The primary outcome was the cumulative live birth rate (CLBR) per stimulated cycle. RESULTS: There were a total of 10,283 ETs (n=5,639 eFET group; n=4,644 fresh group). The freeze-all strategy is associated with improved CLBRs in normal and hyper-responders, but not in suboptimal and poor responders. In Group 1, there were 351 IVF cycles and 387 ETs in total, and the CLBR was 14.3% and 17.7% (p=0.584) for the eFET and fresh group, respectively. In Group 2, there were 2,074 IVF cycles and 2,465 ET in total, and the CLBR was 25.1% and 23.3% (p=0.083) in the eFET and fresh group, respectively. There was a significant difference in the CLBR in Groups 3 and 4, favouring the eFET strategy. In Group 3, 2226 IVF cycles and 3243 ET were performed. The CLBR was 40.5% in the eFET and 36.6% in the fresh group (p<0.001). In Group 4, there were 2547 IVF cycles and 3,188 ET in total, and the CLBR was 52.2% and 47.7% (p<0.001) in the eFET and fresh group, respectively. The number needed to treat to achieve one additional live birth was 25.9 in Group 3 and 22.3 in Group 4. CONCLUSIONS: The implementation of the freeze-all strategy should be individualized. The freeze-all strategy is associated with improved CLBRs in normal and hyper-responders, but not in suboptimal and poor responders.