Calibration
No actuals logged yet. Once a few users hit "Log actual outcome" on a saved plan, this dashboard will populate.
By route
Origin → destination corridors with the most calibration drift. Sorted by |variance|, weak samples (< 3) sink.
By category
Product categories — which goods are we over- or under-estimating?
By origin
Source-country bias — costs of running into surprises in a specific origin.
By destination
Destination-country bias — different VAT/duty regimes show up here.
Reading the numbers. Avg variance % is value-weighted — a single large plan pulls the average more than ten small ones, which is what an operator cares about. Weak rows have fewer than 3 actuals and are statistically too noisy to act on. Significant rows are highlighted in their direction colour (red = over-estimating, green = under-estimating) and have ≥ 3% drift backed by ≥ 3 samples.