Definition
Composite score measuring the degree to which AI-assigned driver-load pairings match the driver's demonstrated proficiency with that specific cargo type, route complexity, and customer relationship history.
The Measurement Gap
Standard dispatch performance metrics evaluate outcomes without examining whether the assignment itself set the driver up for success. The blind metrics include:
Cost per mile: measures economic efficiency of each assignment but never asks whether a lower-cost driver assignment came at the expense of matching a driver experienced with that cargo type, lane, or customer.
On-time delivery rate: captures whether loads arrived on schedule but cannot distinguish between on-time deliveries where the driver was well-matched versus ones where the driver overcame unfamiliarity through extra effort or luck.
Driver turnover rate: signals workforce dissatisfaction broadly but never connects turnover to the pattern of receiving assignments outside a driver's demonstrated strengths, which contributes to frustration and errors.
No legacy metric asks: did the optimizer consider what this driver is actually good at? Dispatch systems optimize for availability, cost, and hours-of-service compliance. Driver proficiency with specific work (cargo handling experience, route familiarity, customer relationship history) exists in institutional knowledge but never enters the assignment algorithm. The result is avoidable damage claims, preventable customer complaints, and driver frustration that shows up in turnover statistics months later.
Decision Supported:
Is the optimizer assigning loads based only on cost and availability, or does it account for the driver's demonstrated capability with this specific work?
Formula
Plain Language
DLPMS scores each driver-load assignment on three dimensions: how experienced the driver is with this cargo type, how familiar the driver is with this route or lane, and whether the driver has a positive history with this customer. Each component is normalized to a 0–1 scale and weighted to produce a composite score. A fleet-wide average reveals whether the dispatch algorithm systematically considers proficiency or ignores it.
Notation
$$DLPMS = \frac{w_1 \times CargoExp + w_2 \times RouteFam + w_3 \times CustHist}{w_1 + w_2 + w_3}$$
Default weights: $w_1 = 0.4$, $w_2 = 0.3$, $w_3 = 0.3$
Variables
| Variable | Description | Computation |
|---|---|---|
| CargoExp | Cargo type experience score | Driver's successful deliveries with this cargo type divided by fleet median for that cargo type, capped at 1.0 |
| RouteFam | Route familiarity score | Driver's completed runs on this lane or zone divided by fleet median for that lane, capped at 1.0 |
| CustHist | Customer history score | 1.0 if driver has served this customer with no complaints; 0.5 if no prior history; 0.0 if prior complaint exists |
| w₁ | Cargo experience weight | Default 0.4: cargo handling is the highest-risk proficiency dimension |
| w₂ | Route familiarity weight | Default 0.3: lane knowledge affects efficiency and safety |
| w₃ | Customer history weight | Default 0.3: relationship history affects service quality outcomes |
All component scores are capped at 1.0 to prevent high-volume drivers from inflating the composite through volume alone.
Proposed formula, pending validation pilots.