The wide use of artificial insemination by dairy farmers has facilitated the development of a multi-billion dollar international market in animal genetics. In the major western dairy producing nations, each country has developed a single index to rank bulls, based on the value of traits they are expected to pass on to their offspring. One of the assumptions behind these indexes is that there is a positive linear relationship between profit (and welfare) with increases in a particular trait, regardless of the farm system. In this paper, it is shown, with examples, that the assumption of linearity is false. More importantly, it is shown that for a combination of reasons, including risk aversion, constraints and other issues, the optimal direction of genetic improvement for New Zealand dairy farmers on an individual and industry level could be quite different. Alternatives to the “one size fits all” index are described.