Freight Budget & Variance Simulator S2

Six-factor PVM bridge · budget build · Monte Carlo risk band · shared assumption registry

Linehaul is modelled as volume × miles per pound × rate per mile and decomposed by sequential substitution into Volume / Mix / Network / Rate, with Fuel and Accessorial separated out. The six factors sum exactly to the actual change — the residual is zero by construction. Sequential substitution is order dependent, so the Shapley symmetric average is shown alongside to expose the bias.

Import actuals

Required
month (YYYY-MM) · volume_lb · linehaul_usd
Recommended — without these the matching factor cannot be split out
miles → isolates Network
fsc_usd → isolates Fuel
accessorial_usd → isolates Accessorial
mode (TL/LTL/PARCEL/IM) · dist_band → isolates Mix

Demo data generator

Months24

Annual volume growth+15%

Annual rate increase+4%

Network gain (miles per lb)−6%

Diesel (base year → compare year)$3.80 → $4.20

LTL → TL shift (mix)5%p

Seed

⚠️ Generated data exists to validate structure, nothing more. Defending a real budget requires real actuals.

Data summary

0 rows

Month	Volume Mlb	Miles k	Linehaul	FSC	Accessorial	Total

Comparison periods

Base

Compare

Decomposition method

Sequential substitution is order dependent. A large gap between the two methods means that factor's attribution is unstable — check it before using the number to allocate credit.

Change in total freight cost

Six-factor decomposition

Factor	Amount	vs base	Control	Shapley gap

Unit cost metrics

Fuel breakdown

Segment contribution (top 10)

Segment	Δ volume	Δ mi/lb	Δ $/mi	Δ cost

Fuel surcharge formula (varies by contract)

Base diesel ($/gal) Step size ($/gal) FSC per step ($/mi)

Now

FSC ($/mi) = max(0, (diesel − base) ÷ step) × FSC per step.
US practice pegs this to the EIA weekly retail diesel index. Replace with the actual contract schedule.

Shared assumption registry (feeds v1 / v3 / S1)

Three tools holding their own copies of the same parameter is how drift starts. They are managed here and exported.

Parameter	Value	Used by

Method documentation

Model — Linehaul_s = Q_s × d_s × r_s
Q = volume (lb), d = miles per lb, r = rate per mile, s = segment (mode × distance band)

Substitution order — Volume → Mix → Network → Rate
L₀=Q₀Σs₀d₀r₀ / L₁=Q₁Σs₀d₀r₀ / L₂=Q₁Σs₁d₀r₀ / L₃=Q₁Σs₁d₁r₀ / L₄=Q₁Σs₁d₁r₁

Volume=L₁−L₀ · Mix=L₂−L₁ · Network=L₃−L₂ · Rate=L₄−L₃ · Fuel=ΔFSC · Accessorial=ΔAccessorial

⚠️ Fix this order and never change it. A different order produces different factor values and breaks comparability with prior years. Consistency matters more here than absolute precision.

PVM bridge waterfall

Monthly freight cost — actual and budget

Unit cost trend ($/cwt)

Monte Carlo budget distribution

Variance contribution — what makes the budget uncertain