Data-Driven Supplement Allocation to Individual Lactating Cows: Precision Feeding on Pasture

In-parlor feeders and electronic cow ID have made it possible to give every cow a different amount of concentrate at every milking. Sales material makes the case in one line: high producers get more, low producers get less, and nobody wastes feed. The agronomy and the economics are less simple. Before you spend money on individual allocation, it helps to know what the published trials found and which data you would actually use.

The idea in plain terms

Under flat-rate feeding, every cow gets the same supplement. Under individual allocation, an amount is set for each cow based on some measured trait such as milk yield, liveweight, body condition, days in milk or pregnancy status. The theory is that cows differ in how much milk they return per unit of feed, so targeting supplement to the cows that respond most should raise herd returns.

On a pasture-based system there is a second wrinkle. Cows already get most of their nutrition from grass, supplement partly replaces grass (substitution), and the response curve flattens. Penn State's summary on Farms.com puts expected substitution at roughly 0.4 to 0.6 and shows milk response per pound of concentrate falling as the amount rises. A flat curve means any gain from shifting feed between cows is small unless the cows differ a lot.

What the DairyNZ review found

A DairyNZ technical series document on individualised feeding pulled together international work and local modeling. Key points:

• In-shed feeding systems were in about a third of New Zealand dairies at the time, with roughly a quarter of those able to feed cows individually.
• An Australian study reported about a 3% rise in milk yield and 7% in fat and protein yield with individual feeding, though maize silage in the diet may have influenced that.
• An Irish study found no significant difference in intake, milk yield, milk solids or energy balance between flat-rate and milk-yield-based feeding.
• A UK grazing study found little effect when pasture was relatively unrestricted.
• Whole-farm modeling compared 170 individualised strategies per system against flat-rate feeding at the same total supplement and found none that beat flat-rate on operating profit.
• A four-week on-farm experiment found no herd-level difference in production, liveweight or condition score.

The report's conclusion was that there was no evidence individualised feeding improved profitability or the milk solids response to supplement compared with flat-rate feeding in pasture systems, and that the systems cost more to install and manage. It did note operational benefits such as giving extra feed to thin cows without managing separate mobs.

That conclusion comes from New Zealand's low-input, low-supplement system, where a milk-to-feed price ratio is much narrower than in the US. The Penn State summary notes that at ratios near 2.0, feeding 16 to 20 lb to high-merit cows can make sense, so a heavier-supplement herd could see different economics.

What a precision strategy looks like in a modern trial

A more recent trial from France's INRAE, published in March 2025, takes the idea a step further by adjusting concentrate based on each cow's response rather than on simple rules. In a four-month study of 40 multiparous Holsteins, matched pairs followed either a standard fixed ration or a precision strategy with weekly concentrate adjustments based on body weight gain. The summary on INRAE's website reports no significant difference in milk yield, intakes or overall weight gain between groups, so the precision group performed similarly while the headline is that it could do so with less concentrate overall. Blood metabolites (glucose, BHB and non-esterified fatty acids) sorted cows into two metabolic clusters with about 90% accuracy, and simpler measures such as body weight gain and feed intake predicted the clusters with up to 70% accuracy. The authors note that long-term effects still need study.

It is a small trial, and it shows that some cows can be fed less without losing milk, not that every herd will profit from the equipment.

What data actually drives a good allocation

If you do decide to try it, the data feeding the decision matters more than the hardware. Useful inputs include:

• Milk yield and components. Yield is the most common basis, and it is cheap to get from your parlor or DHI test. See our guide to reading a DHI record.
• Body condition and weight change. The thin cow is the one that most plausibly justifies extra feed. It is also a measurement that takes discipline, not just a sensor.
• Days in milk and pregnancy status. Early lactation cows have the most to gain.
• Health flags. A cow with a metabolic problem or a lameness issue is not a candidate for more concentrate to fix it. See the lameness guide.

Questions to answer before buying

1. How much do my cows really differ? If yield and condition are fairly uniform, there is little to reallocate.
2. What is my milk-to-feed ratio? It decides whether extra concentrate pays at all.
3. Is pasture allowance generous or tight? The UK work found little effect when pasture was relatively unrestricted.
4. Can I get most of the benefit by grouping? A fresh pen and a late-lactation pen may deliver most of the value for far less cost.
5. Who will maintain it? The Dairy NZ review warned of cost and complexity.

Bottom line

The research does not support a blanket case for individual concentrate allocation on pasture. Flat-rate feeding held its own in New Zealand modeling and trials, and the newer French work suggests the gain is more about saving feed than raising milk. The strongest practical case is targeted help for thin or early-lactation cows. Measure the spread in your herd first, price the equipment against that, and consider simpler grouping before buying hardware. For more on herd data tools see our section on the agriculture directory, and browse farm.fm's dairy pages.