Case study · Logistics

A parcel carrier: volumes forecasted down to postcode.

A proof of concept that tested whether MyIO can forecast parcel volumes across a national logistics network.

Logistics warehouse aisle
45Mparcels a year in the analysed scope
3.9%average global forecast error
80%of forecast errors below 5%

The task

A European parcel carrier needed to know whether MyIO could forecast parcel volumes accurately across its network. The proof of concept forecast the first quarter of 2023 retrospectively and checked accuracy at every level of detail.

What we built

A hierarchical, probabilistic forecasting model covering the whole structure of the network: global, countries, regions and individual postcode areas, with a four-week lead time. Model performance was monitored against actual volumes.

The result

  • 3.9% average global forecast error
  • 80% of forecast errors below 5%
  • Forecasts consistent from global level down to postcode
Client
European parcel carrier (name under NDA)
Industry
Parcel logistics
Product
MyIO
Scope
Retrospective forecast of Q1 2023

Tell us what you want to solve. We’ll show you where AI pays off.