How we measure route optimisation results
The method Route4Green applies in every Pilot and every published figure: three comparison lanes, volume normalisation, the metric set and when a metric is withheld, and seven honesty rules.
Why one consistent method
Route optimisation software can always produce a percentage. The question worth asking is what that percentage is divided by, compared against, and measured over. This document describes how Route4Green measures results in every Pilot and every published figure, so that a customer's own analyst can re-check each number and so that our figures stay comparable over time.
Three comparison lanes
Every efficiency claim names which two of three lanes it compares: the baseline plan (the routes the operation uses today), the optimised plan (what the system produced for the same orders), and what was actually driven (odometer or telematics). Plan versus plan is a valid measurement but not field evidence; the minimum evidence for any field claim is at least one real shift, with real vehicles, with odometer readings. Fixed-route standard kilometres are usually padded, so baseline km has to be named for what it is.
Volume normalisation, with a worked example
Existing routes usually also carry cargo for points outside the trial scope, so the two sides of a comparison serve different volumes. The document sets the order of precedence: strip out-of-scope cargo when per-stop data exists, otherwise compare only against baseline routes that serve in-scope points, and otherwise normalise every metric by volume with the factor printed in the table. The worked example is the VNPost Pilot in Ho Chi Minh City, July 2026, which shows why the raw totals are never the headline.
The metric set, and when a metric is withheld
Ten metrics, each with its formula and the condition under which it is not reported: trips, distance, cargo per route, load utilisation, working hours, cost, emissions, time-window compliance, rejected records and run size. A metric whose input data is missing is reported as not measured; we do not estimate in its place.
Seven honesty rules and what a measurement does not conclude
Seven rules travel with every report: every percentage carries its denominator, cargo per route is never called load utilisation, hour models are never mixed, cost needs the same components on both sides or is dropped, emissions if reported are computed from normalised km with a sourced factor, no feasibility claim before per-leg load is checked, and numbers that do not qualify for publication stay in the appendix rather than being deleted. The document closes with why Route4Green publishes no emissions figure of its own today, and three things a measurement does not conclude.