Can I trust the data provided by DigiFarmz in producing regions where there are no protocols?

Agricultural experimentation is always based on sampling, across all areas.

Genetics:
The VCU (used for cultivar registration) and final germplasm trials are carried out in representative regions, but not in every municipality of the country.

Soil and nutrition (chemical and physical):
Sampling is performed at specific points in the field (grid sampling), but never across every square meter of a given area.

Climate and physiology:
Modeling in these areas is based on experimental data that is extrapolated (following well-defined principles and criteria), giving rise to laws. For example, the FAO’s estimated yield model is used worldwide, based on trials originally developed in Europe and the United States.

Crop protection products:
The Ministry of Agriculture requires three technical reports obtained in three different experimental environments (within the same year) or two experimental environments (across two years). However, these experiments are not conducted in every location. The industry, in the final pre-launch phase of a product, conducts an important group of trials, but not across every municipality in a country.

The key point is that, with the right selection of parameters, weights, and metrics, it is entirely possible to obtain reliable efficacy data from remotely conducted experiments. Naturally, as DigiFarmz expands its coverage and benefits from crowdsourced data, accuracy will continue to improve.

Currently, the market is supplied with information whose accuracy variation may exceed 20%.

But the fundamental point is this: the control achieved with two products showing 50% and 80% efficacy can result in a similar level of disease control, provided that the timing and intervals between applications are adjusted according to those efficacy values. This is precisely the principle behind DigiFarmz’s modeling—delivering high control effectiveness even when the available factors are not proportionally high.