Pilot
Compare forecasts with how you plan today
Select a crop, a region and a historical period. We generate forecasts at several lead times before harvest and compare them with the actual outcome and your existing estimate.
Process
Four steps
- 01
Select crop, region and period
Choose a crop, a region and a historical period where outcomes are known.
- 02
Forecast at several lead times
Generate forecasts at several lead times before harvest, as if run in that season.
- 03
Compare
Compare with the actual outcome, the customer's existing estimate and any available baseline.
- 04
Judge planning value
Measure whether the additional signal has meaningful planning value, and at which lead time.
Measures
What a pilot can measure
- Forecast error by lead time
- Lead time gained versus the current method
- Regional coverage
- Planning decisions the forecast could have informed
Measures are agreed per pilot. Results are published only once a pilot has been completed and agreed with the client.