The objective of this study was to review some of the latest evidence on genetic variation in feed intake and feed utilization and to determine how this variation might be used. The most important sources of genetic variation in gross efficiency are likely to be the quantities of feed eaten and used for yield or maintenance and the extent to which body tissue is mobilized. Accounting for just one of these components when selection is for improved feed efficiency might result in undesirable genetic changes. For example, in an ad libitum feeding system, the heritability of body condition score is reported to be 0.43 for heifers; genetic correlations of body condition score with milk production and live weight were -0.46 and 0.66, respectively. Also, the genetic correlation between milk yield and live weight depends on lactation stage. For example, over the first 26 wk of lactation, this correlation was reported to be -0.09, but, after genetic adjustment for body condition score, the correlation was 0.29. When economic values are being derived, energy norms or genetic correlations can be used, and double counting of the feed costs needs to be avoided. An index that contained linear type traits, however, gave high accuracy of selection. Hence, although there appears to be great potential to improve economic efficiency by selecting for feed intake and live weight or by possible indicator traits, there is still uncertainty about some of the genetic parameters, especially among traits related to health, reproduction, and energy balance.