Logged before. Scored after.
We build forecasts that are written down before the outcome is known and scored against what happened, so you know how far to trust them.
A forecast nobody scores cannot be trusted.
Most forecasts are judged by how convincing they sound. When they miss, nobody checks, and the next one sounds just as convincing.
Every forecast we publish is logged with a timestamp before the outcome, and scored when the outcome is known. CERES does this every week for food crises in 43 countries, and the same discipline applies to demand, prices or load.
What we deliver.
- Build the forecast from your own data and the public sources around it.
- Log every forecast with a timestamp before the outcome is known.
- Score it against what happened, and show the record.
- State the uncertainty, instead of hiding it.
How it works.
- Define
- We agree on what is forecast, how far ahead and how it will be scored.
- Log
- Every forecast is written down before the outcome.
- Score
- The record is kept and shown, the misses with the hits.
Data and industrial intelligence
More in this area.
All solutions- Root cause in process dataWe rank what drives a loss, a quality problem or downtime, from the data you already record, and trace every cause back to the raw signal.
- Independent monitoringWhere a question reaches beyond your own sensors, we confirm it from satellites and public records: ground motion, infrastructure, forests and activity at sea.
- Data pipelinesWe bring data from many systems and sources into one place, keep it up to date and record where every value came from.
- AI in your systemsWe build assistants and drafting into the systems your people already use, with the data kept in Europe and a person approving what goes out.
