Predict & Planning
Yield forecasts — and how wrong they were.
Every vendor will forecast your yield. We publish the model's accuracy back to you, cycle after cycle.
Yield prediction is the easiest claim in agricultural software to make and the hardest to check. A number appears; the harvest happens; nobody goes back and compares them.
Digital Cultivation forecasts growth from the crop recipe and the conditions the plants have actually experienced — and then, when the harvest is logged, feeds the real weight back into the model. Which means the platform knows how accurate it has been, per crop, and shows you.
You get the forecast. You also get the error bar, honestly reported. If the model is bad at cilantro, you will find out from us rather than from a missed order.
Questions
The things people actually ask.
How much history do you need before it is useful?
The forecast works from the recipe on day one. Accuracy improves as harvests feed back in, which is exactly why the accuracy figure is shown rather than hidden — you can see for yourself when it has earned your trust.
What does "model accuracy" actually mean here?
Predicted yield against actual harvested weight, per crop, reported as mean error along with the bias — so you can see not just how wrong it was, but whether it is consistently optimistic.
Does it use our vision data?
Where Crop Health is enabled, canopy measurement gives the model a real signal on how the crop is developing instead of relying on elapsed days alone.
See Predict & Planning on a real farm.
We'll walk an order all the way through to its margin — on live farm data, not a slide.