The article says that implantation fails in humans 10-40% of the time. Your point is still valid, but the scale in reality is very significant.