The objective was to use ovulation synchronization with timed artificial insemination (TAI) to evaluate the effect of timing of artificial insemination (AI) with frozen sex-sorted sperm on fertility performance in pasture-based compact calving herds. Ejaculates from 3 Holstein-Friesian bulls were split and processed to provide frozen sex-sorted sperm (SS) at 4 × 10 6 sperm per straw, and frozen conventional sperm at 15 × 10 6 sperm per straw (CONV). A modified Progesterone-Ovsynch protocol was used for estrous synchronization, with TAI occurring 16 h after the second GnRH injection for cows assigned to CONV, and either 16 h (SS-16) or 22 h (SS-22) for cows assigned to SS. Pregnancy diagnosis was conducted by transrectal ultrasound scanning of the uterus 35 to 40 d after TAI (n = 2,175 records available for analysis). Generalized linear mixed models were used to examine the effects of treatment on pregnancy per artificial insemination (P/AI). Fixed effects included treatment (n = 3), bull (n = 3), treatment by bull interaction, parity (n = 4), days-in-milk category (n = 3), and treatment by days-in-milk category, with herd (n = 24) included as a random effect. Pregnancy per AI was greater for CONV compared with both SS-16 and SS-22 (61.1%, 49.0%, and 51.3%, respectively), and the SS treatments did not differ from each other (relative P/AI for SS-16 and SS-22 vs. CONV were 80.2% and 84.0%, respectively). There were significant bull and treatment by bull interaction effects. Additional analysis was undertaken using a model that included herd as a fixed effect. This analysis identified marked herd-to-herd variation (within-herd relative P/AI for the combined SS treatments vs. CONV ranged from 48-121%). The tertile of herds with the best performance achieved a mean relative P/AI of 100% (range = 91-121%), indicating that P/AI equivalent to CONV is achievable with SS. Conversely, the tertile of herds with the poorest performance achieved a mean relative P/AI of 67% (range = 48-77%). We found that SS resulted in poorer overall P/AI compared with CONV sperm regardless of timing of AI. Marked variation existed between herds; however, one-third of herds achieved P/AI results equal to CONV. Identification of factors responsible for the large herd-to-herd variation in P/AI with SS, and development of strategies to reduce this variation, warrant further research.