Benchmarking your yields without kidding yourself

There are three things to benchmark a yield against, in rising order of usefulness: the national average, your own field history, and genuinely comparable farms. Most benchmarking effort goes on the first and third, which are context and gossip respectively. Nearly all the decisions come out of the second — and it's the only one entirely within your power to get right.

The catch with all three is the same: the comparison is only as honest as the numbers going into it, and yield figures have several well-worn ways of flattering their owner.

The three comparisons

Comparison	Good for	The trap
National averages	Context; sanity-checking a season	Mixes all land, all rotations, all luck
Your own field history	Actual decisions	Only exists if you've kept it
Other farms	Motivation, occasionally insight	Pub figures are survivor-biased

National averages tell you what kind of season it was, nationally — useful for deciding whether your disappointing year was your fault or everyone's. The UK crop yield benchmarks are the reference here. But the national figure averages first wheats on fen silt with third cereals on sand, so "above average" is a low bar on good land and an impossible one on thin land. Context, not a target.

Your own field history is the benchmark that changes behaviour: this field, this rotation slot, the last three to five seasons. It answers the questions that actually cost or make money — does second wheat pay here, did the drainage work, is that variety earning its seed cost — and it's the comparison why one field yields more than another depends on. It requires records, which is the entire difficulty. Field histories covers what to keep.

Other farms would be the gold standard if the data existed: same soil, same rainfall, same rotation, honestly measured. It mostly doesn't. Discussion-group figures come closest. Pub figures are worth what you paid for them — the neighbour's 11 t/ha field gets mentioned; his 6.5 t/ha field, cut in the dark and hauled wet, somehow never comes up. Farms that had a shocker don't volunteer numbers at all, so the "local average" you hear is an average of the willing.

The honesty rules

Three rules keep a benchmark from quietly flattering you. Break any one of them and the comparison is decoration.

Same moisture basis. A hundred tonnes at 18% moisture is not a hundred tonnes at 14.5% — the difference is about 4 tonnes of water being counted as wheat. Adjust everything to one standard moisture before comparing fields, seasons or farms. Wet-year yields compared raw against dry-year yields will "prove" whatever the wet year wants.

Cropped area, not registered area. Divide by what was drilled, not the field-name area on the maps. Margins and stewardship strips commonly shave 5–10% off the cropped area, and a 5% area error is a permanent 5% yield error — always in the direction of making the field look worse. Measuring field area covers getting this right once.

Weighed tonnage, not the yield monitor's opinion. Uncalibrated yield monitors are commonly 3–5% or more adrift; calibrated ones still run 1–3%. Monitors are excellent at relative patterns within a field and unreliable as the season's official total. The whole-field number should come from weighed loads — which, as a bonus, is the number the grain trade will eventually hold you to anyway. The field yield calculator turns weighed tonnes and cropped area into a clean t/ha.

Cost per tonne beats yield per hectare

Yield per hectare is the number farmers compare; cost per tonne is the number that decides cropping. A field doing 8 t/ha on £1,100/ha of growing costs produced wheat at about £137/t. A field doing 7 t/ha on the same spend produced it at £157/t — £20/t worse, which at recent feed wheat values is most of the margin.

This is where benchmarking earns its keep. A field can be below the farm's average yield and still be its cheapest wheat, because it gets there on less spend; another can top the yield league while producing the most expensive grain on the farm. Rank the fields by cost per tonne once and the cropping conversation changes — some of the "good" fields stop looking clever, and one or two quiet performers turn out to be carrying the place.

You can't do that ranking without honest per-field tonnages. Which is the point.

What three seasons of honest records buys you

One season, on its own, proves little — any field can be excused or flattered by one year's weather. Three seasons of moisture-adjusted, weighed, cropped-area yields is different in kind:

• The persistently poor fields separate from the unlucky ones, and each gets a diagnosis and a decision — drain it, lime it, or accept it and cut its costs.
• Rotation questions get real answers: the true second-wheat penalty on your land, in your figures, not a conference slide's.
• Budgets stop being hopeful. Forward selling against a field-by-field five-year average is a different exercise from forward selling against optimism.

Keeping the record is a discipline problem during harvest more than an analytical one afterwards — loads want recording against fields as they happen, not reconstructed in November. That capture problem is what Harvestt was built for, but however it's done, done is what matters.

The benchmark that improves a farm is the one against its own past, measured the same way every year. Everything else is conversation.

Frequently Asked Questions

What should I benchmark my wheat yield against?

Primarily your own field-by-field history over the last three to five seasons, at a standard moisture — that's the comparison that drives real decisions. National averages (around 8 t/ha for winter wheat) are useful context for judging the season, and other farms' figures are interesting but rarely comparable or complete.

Why do my yields look worse than the yield monitor said?

Usually some mix of monitor error and area error. Uncalibrated monitors run 3–5% or more out, and dividing weighed tonnage by registered rather than cropped area understates yield by another 5–10% on fields with margins. Weighed loads over drilled area is the honest figure.

Is cost per tonne more important than yield?

For cropping decisions, yes. Yield per hectare ignores what the yield cost to grow — a 7 t/ha field on modest spend can produce cheaper grain than a 9 t/ha field grown expensively, and cheap grain is what survives a poor price year. Use yield to diagnose fields and cost per tonne to choose between options.

How many years of yield data do I need to benchmark properly?

Three seasons is the practical minimum — it usually spans wet and dry years, so persistent differences between fields separate from weather noise. Five is better for budgeting. One season is a data point, not a benchmark.