How do I run an on-farm feed trial?
Split groups, one change, clear records
ELKHORN, Wisconsin — Run an on-farm feed trial by changing one thing, comparing it with a matched control at the same time, measuring the same outcomes in both groups and keeping honest records. If you can only manage one pen, use a switchback design where the same group alternates between the two diets. Keep it simple, keep it fair and don't read too much into a small herd's results.
On-farm trials won't give university-level certainty, but they can tell you whether a change is worth a closer look, and they can save you from buying something that does nothing.
Pick one question
Decide what you want to learn before you start. "Does this additive raise milk?" is a question. "Let's try a few things and see" isn't. Choose one change: a product, a forage, a mineral, a feeding time. Changing several things at once makes the result impossible to interpret.
Then decide which outcomes you'll measure. Common ones are milk yield, fat and protein, dry matter intake, body condition, health events and feed cost. Decide the minimum difference that would make the change worth paying for. If the product costs a given amount per cow per day, calculate how much milk it needs to add just to break even.
Choose a design that fits your herd
If you have two similar pens, assign one as control and one as treatment, then swap them after a period so each pen gets both diets. This reduces the effect of pen differences. If you have one group, use a switchback: feed diet A, then diet B, then back to A, and compare the periods. Each period needs to be long enough for cows to adjust, which your nutritionist can help set.
If you can split a pen, match cows by stage of lactation, parity and milk yield, then assign them to each side. Make the assignment random, not by picking cows you like. Keep the groups on the same schedule, same bedding and same milking time.
Resources like the Wisconsin Extension farm pages and the Sustainable Agriculture Research and Education program have guidance on farmer-led research design, and many offer templates for record sheets.
Collect clean records
Write down:
- Start and end dates for each period
- Exact diet, including any forage changes during the trial
- Animals in each group
- Milk weights, components and any somatic cell count data
- Feed offered and refused, if you can
- Health events and treatments
- Weather or anything unusual
Weigh the feed. Don't guess. If you can't measure dry matter intake by pen, at least track the amount loaded in the mixer and refusals. Test forage during the trial, because a change in silage dry matter can look like a treatment effect.
Keep the rest of the farm steady
Avoid starting a trial the week you open a new silage bunker or change hay lots. Keep feeding times, bunk management and milking routines the same. Have the same people doing the work. Tell the milkers and loader operators what you're doing, because a quiet change by someone trying to help can wreck a month of data.
Also run the trial long enough. Short trials catch only the first reaction, and cows can respond differently after a few weeks. Your nutritionist or extension dairy specialist can help set a sensible length for the question you're asking.
Read the result with humility
When it's done, compare the groups. Look at average differences and at how consistent they were. If one group's milk jumped on a day when something else changed, discount that point. A tiny difference in a small group is probably noise. A bigger difference that appears in both pens and both periods is more believable.
Consider the economics. Subtract the extra feed cost from the extra income, and see whether the change pays. Tools from the Center for Farm Financial Management can help with the budgeting. Then talk to your nutritionist, and if the result is promising, repeat the trial, since a second run is how you separate real effects from luck.
- State the question and the one change.
- Choose a control and a design: split pens, swap pens or switchback.
- Decide outcomes and the break-even gain.
- Set a start date when nothing else is changing.
- Record feed, milk and health consistently.
- Test forage during the trial.
- Review the data with your nutritionist.
An on-farm trial isn't a research paper. It's a careful way to avoid fooling yourself, which is harder than it sounds and cheaper than most mistakes.