How Growgoyle Analyzes a Cultivation Run

Four rules govern every analysis.

Start with the batch record

Evidence is selected by facility, room, flowering batch, phase, and date. A reading without crop context is not treated as a conclusion.

Separate fact from inference

Measured differences are stated as facts. Explanations are framed as possible contributors and must include a plausible mechanism.

Expose uncertainty

Confidence, confounders, missing prerequisites, and data-quality limitations remain visible instead of being hidden behind a confident paragraph.

Prefer reversible tests

The safest useful output is usually one bounded change, with measurements to monitor and an explicit condition for stopping or reverting.

What the system can consider

Available context varies by facility and batch. Growgoyle uses only data that exists and reports important gaps.

Environment and configured targets Lighting and photoperiod Irrigation, feed, and runoff Cultivar and phase Grower notes and observations Recorded interventions Incidents and equipment context Harvest, quality, and lab results Comparable prior runs

How to read a Growgoyle finding

Example structureA finding is an evidence chain, not a verdict.

Observed result

The measured outcome and its relevant baseline or comparison.

Possible contributor

The operational or biological hypothesis most worth investigating.

Supporting evidence

Specific readings, timing, observations, and comparable-run differences that strengthen the hypothesis.

Limits and confounders

Other changes, missing data, quality tradeoffs, or conditions that weaken the conclusion.

Next-run experiment

The smallest measurable and reversible change supported by the available evidence.

Monitor and revert

Crop responses and thresholds that tell the operator whether to continue, adjust, or stop.

What Growgoyle will not claim

It will not claim that observational correlation proves a single cause.
It will not confirm a pathogen, pest, or deficiency from a photo when the evidence requires scouting, microscopy, or laboratory testing.
It will not guarantee yield, potency, quality, or financial improvement.
It will not replace the operator's judgment, company SOPs, label requirements, worker-safety controls, or local regulations.

When the record is not strong enough: Growgoyle should say insufficient data, identify the missing measurement, and recommend how to close the gap. A smaller honest answer is more useful than a confident guess.

Inspect a completed-run analysis.

See how measurements, opportunities, missing data, and next-run suggestions appear in the product.

Open the example analysis