Feed efficiency in livestock: what the ratios measure and what they hide
Published 30 July 2026 · Updated 1 October 2026
Feed conversion ratio is the number every producer knows and the one animal nutritionists trust least. It divides feed eaten by weight gained, so it carries every difference in water content, maintenance cost, tissue composition and feed energy density in a single figure, and any of those can move it without the animal becoming more or less efficient in any physiological sense.
FCR and the problems built into it
A modern broiler reaches 2.5 kg at about 38 days with an FCR near 1.5. A grower-finisher pig from 30 to 115 kg converts at about 2.5 to 2.8. A feedlot steer gains at roughly 6 to 7 kg of dry matter per kg of liveweight gain. Farmed Atlantic salmon sits near 1.1 to 1.3. Ranked on FCR, salmon beat broilers, broilers beat pigs, and pigs beat cattle by a wide margin. The ranking is real but says much less than it seems to.
Fish gain more weight per unit of feed in part because they are ectotherms and spend nothing on holding body temperature; broilers grow quickly and are slaughtered young, when most of the gain is lean tissue with a high water content; the steer is slaughtered at the end of a long fattening phase when gain is mostly fat, carrying 15 to 20 MJ per kg compared with 8 to 10 for broiler gain. The energy in a kilogram of gain differs by a factor of two across species and by a factor of nearly two within one species across body weight. Dry matter conversion helps (many beef systems quote gain:feed on a dry matter basis for this reason), but it does not remove the energy density problem.
Maintenance enters as a fixed daily cost. A slow-growing animal pays more maintenance per kilogram of gain than a fast one, so FCR improves with growth rate in nearly every species, independently of any change in metabolism. Selecting for growth improves FCR as a side effect. Hence the push for a measure that holds growth and size constant.
Residual feed intake
Residual feed intake (RFI), introduced by Koch and colleagues in 1963 for beef cattle, is the difference between actual feed intake and the intake predicted from body weight and gain. A negative RFI describes an animal that ate less than expected for its size and performance. It is uncorrelated with growth rate and body weight by construction, which is the point of it. Heritability is moderate, in the range of 0.3 to 0.4 in cattle and pigs, and low-RFI cattle eat about 10 per cent less feed for the same gain.
What RFI hides is the source of the variation. Studies partition the differences among animals into digestion, heat increment of feeding, protein turnover, body composition, activity, and the cost of tissue maintenance, with no one component dominating. In pigs, low-RFI animals tend to be leaner, which would be consistent with the net energy cost of protein versus fat deposition; in cattle, differences in feeding behaviour and fasting heat production account for a share. RFI is also a measurement of a particular animal on a particular diet during a particular test period, normally 70 days, and the ranking changes if the diet changes. Feeding stations that record individual intake are required, which restricts the method to research herds and breeding programmes.
Partial efficiencies of ME
The physiological alternative is to break efficiency into its steps. Metabolisable energy is used for maintenance with efficiency km and for gain with efficiency kg. In the ARC (1980) framework both depend on the metabolisability of the diet, qm, which is ME divided by gross energy: km = 0.35 qm + 0.503, and kg = 0.78 qm + 0.006. At qm = 0.6 this gives a km near 0.71 and a kg of about 0.47. At qm = 0.7, kg rises to 0.55. The efficiency of gain is always lower than the efficiency of maintenance, since depositing protein and fat costs more than burning them, and the gap widens on poor feeds.
The practical interpretation is that the same 1 MJ of ME is worth more or less depending on what the animal does with it. A cow in lactation uses ME for milk at kl near 0.64, steers for gain at 0.3 to 0.5, and pigs on net energy systems are scored more directly, since the NE value is the ME value after the heat increment has been subtracted. Comparison across species then reduces to energy retained per unit of ME eaten, which is a figure that includes the product (milk, egg, lean, fat) in joules and not in kilograms.
What to compare
If two systems must be compared, they should be compared on edible protein or energy output per unit of edible input, or on whole-chain figures including the breeding herd. On those measures the ranking changes: broilers and eggs are at the top, pork and milk are in the middle, beef is low, and the differences between species shrink or reverse when ruminants are credited with fibrous feed that people and monogastrics cannot use.
Within a species and within a farm, FCR remains an adequate monitoring tool, provided that days on feed, slaughter weight, mortality and feed energy are held constant. For genetic comparison use RFI; for ration comparison use gain per unit of ME or NE. The two ratios answer different questions, and quoting the wrong one at the wrong meeting is how the same data end up with two contradictory conclusions.