HARVENTISFor the first season at Elite Harvest we wrote the harvest down the way every polyhouse we had visited did it: one number per day, in a notebook that lived near the packing table. 380 kilos on Monday, 410 on Tuesday. It felt like record-keeping. It was really just a diary.
The problem showed up about six weeks into the cycle. The daily totals were sliding, slowly, and we could not say why. Was it the whole crop tiring? A few beds with a root problem? A worker picking too early at one end of the house? Every answer was a guess, because the notebook had thrown away the one thing that would have told us: where each kilo came from.
What a bed-wise record actually gives you
We run around 10,800 plants across 56 beds. When the harvest is recorded per bed — even roughly, a crate count per bed per pick — three things become visible that a daily total hides completely.
Weak beds. Within a week it was obvious that a handful of beds at one end were producing 30 to 40 percent less than the rest. They sat near a vent where the drip pressure dropped. Nobody had noticed because nobody was comparing beds; the daily total had simply absorbed the shortfall.
Harvester consistency. When each entry carries the harvester’s name, you see who picks to grade and who picks for weight. That is not about blame. It is about knowing who to put on which rows when a buyer is strict about size.
The real curve of the crop. Cucumber yield rises, peaks and tails off. With bed-wise data the peak is sharp and visible, and you can plan the next sowing against it instead of against a feeling.
The decision it changed
The most expensive decision in a polyhouse is when to pull a crop and replant. Pull too early and you leave fruit in the house; too late and you are paying wages and inputs for kilos that are not coming. With daily totals we argued about it. With bed-wise yield per plant, we could see the point where the bottom third of beds had fallen below the cost of harvesting them, and the top third were still worth keeping. We pulled the weak beds first and staggered the replant. The house never went fully empty, and the cash never fully stopped.
Why nobody does this with a notebook
Bed-wise recording fails on paper for a simple reason: it is too many lines. Fifty-six beds, two grades, a harvester name and a time — that is a page a day that nobody wants to write standing in 38 degrees. So it gets summarised at the packing table, and the detail is gone.
The fix is not discipline. It is making the entry cheaper than the summary. On a phone the harvester taps the bed, taps the grade, types the weight, and is done in under ten seconds, before the crate is even stacked. If the signal is bad under the plastic, the entry waits on the phone and goes up later. The farm office sees it the moment it syncs.
That is the whole reason Harventis records harvest the way it does. Not because bed-level data is clever, but because it is the only level at which the numbers tell you what to do next.
If you want to start tomorrow
You do not need software to begin. Number your beds. Put a laminated card on each one. Ask harvesters to call out the bed number when they hand over a crate, and have one person at the packing table write the bed and the weight. In two weeks you will know which beds are carrying your farm. After that, you will not want to go back — and you will be ready for a system that does the writing for you.