Why You Can't Quote a Freight Offer in Your Head

A freight-board offer pops up. Rotterdam → Madrid, 1 850 km, €1.18 per kilometer. You have two minutes before someone else takes it. The dispatcher next to you accepted what looked like a similar route at €1.10 last week and made money. You glance, calculate roughly, accept.

Three weeks later in the month-end P&L, this trip is somehow worse than the one your colleague did at the lower rate. Same distance bracket, same destination country, similar tractor-trailer combination. What changed.

Five things, none of which the in-head calculation can hold at the speed of a freight-board decision.

1. The per-country price matrix is too big to memorize

Every EU country prices road freight differently, and "differently" means structurally different rules, not just different numbers.

Germany uses Maut indexed to CO2 class, weight band and axle count, with the schedule reworked in late 2023 to push older tractors into a higher tier. France runs péages on operator-specific schedules (Vinci, ASF, APRR each price independently) with vehicle classes that do not map 1:1 onto German ones. Spain charges only on AP autopistas, with vacíos vs cargados differentiation in some operators. Italy stacks gestori privati rates on top of base schedules. Switzerland's LSVA is gross weight × kilometres × emission class, with refund mechanisms most carriers do not bother claiming. Austria sells GO-Boxes priced per kilometer and axle. Poland switched from manual viaTOLL to e-TOLL in 2021. Czech Republic and Slovakia run their own e-mýto. Hungary, Slovenia, Croatia have variants again.

Fuel adds another layer. Country-average diesel prices spread 20–30 % across the EU at any given moment. Excise-tax refund schemes apply selectively (some operations qualify, some are below the volume threshold to make it worth filing). Your real per-litre price depends on your fuel-card discount and where on the network you actually buy, not the country average.

Nobody carries the full matrix in their head accurately. The dispatchers who appear to "just know" the rates have memorized maybe four or five corridors well and approximate the rest. The cost of being wrong on Switzerland for one trip is around €200. Multiplied across a year of approximate quoting, this adds up to a number that nobody enjoys finding in the year-end review.

2. Route composition compounds the variance

Even within a single country pair, the same kilometres can cost differently depending on which routes pass through which jurisdictions.

Berlin → Madrid is roughly 1 850 km via France. The same destination via Italy and southern France is closer to 2 300 km, but the toll mix is fundamentally different. Even within the French-corridor version there are two viable variants (northern A4 vs central A6/A7), and the toll difference between them on a typical 40-tonne 4-axle tractor-trailer is €60–80 per trip. Same distance bracket, different operator stack, different total.

Cross-corridor comparisons get worse. Berlin → Milan and Hamburg → Sofia are similar in total distance but pass through entirely different country mixes (DE + AT + IT vs DE + CZ + SK + HU + RO + BG). Different rate schedules, different e-tolling systems, different vehicle-class definitions, different fuel pricing regimes. The "average toll per kilometre" you computed from last quarter's books is dominated by whichever corridors fill your books, and is silently wrong for any corridor you do not run often.

The 1 850 km of the Rotterdam → Madrid offer above is not the same 1 850 km as the route the colleague ran last week. The book of memorized rates does not zoom that far in.

3. Fixed costs amortize per hour, not per kilometer

This is the one most experienced dispatchers feel without articulating, and it is the place where the €/km comparison structurally cannot be right.

Take two offers, both 700 km, both reasonable destinations.

Offer A: €1.30/km. The shipper has loading window 08:00–11:00 and a delivery window 14:00–18:00 the next day. The truck spends 3 hours at pickup and 4 hours at delivery on top of the drive itself. Actual truck-and-driver time tied up: about 16 hours.

Offer B: €1.10/km. Both ends run through a 3PL with rapid turnaround. 1 hour pickup, 1 hour delivery, plus the drive. Total: about 11 hours.

Same 700 km, so variable costs (fuel, tolls) are essentially identical between A and B. But fixed costs amortise against time, not against distance. A European own-fleet tractor-trailer with own driver typically carries somewhere around €20–30 of fixed cost per hour (driver wage with social charges + leasing + insurance + depreciation + vehicle-attributed overhead, divided by working hours per year). Call it €25/h for round numbers.

Offer A wins, but by a narrower margin than the €/km difference implied. Now imagine Offer A actually waits 6 hours at delivery instead of 4, which is a typical industry slip. Another 2 hours × €25 = €50 evaporated. A becomes −€30, an outright loss. B is still €5, marginal but positive. The €0.20/km premium did not pay for the extra time.

This is what experienced dispatchers feel when they say "that €1.10 load worked out better than the €1.30 one." It is not gut feel. It is a real calculation. The €/km comparison cannot make it because €/km treats time as if it scaled with distance, which it does not. The kilometres are paid for by the customer at the agreed rate. The hours are paid for by you, regardless of how the customer prices the job.

4. Time costs cascade non-linearly at certain thresholds

Within the same load, the gap between 9 hours and 11 hours of drive time is mostly linear: +2 hours × hourly fixed cost. But the gap between 11 hours and 13 hours crosses the daily driving limit, which triggers an 11-hour mandatory rest period. The trip now finishes 13 hours later in clock time, not 2 — or you double-crew it, which is a different and real cost. Either way, 2 extra hours of driving turned into 13 extra hours of amortisation, plus possibly a second driver.

Similar step functions live elsewhere. Missing a ferry slot by an hour can mean waiting 6 for the next one. Crossing into a second day of a time-based vignette doubles the day count for the duration of the trip. Loading delays that push delivery into the next morning may trigger an overnight stay with parking and per-diem.

These are step costs, not linear costs, and the step points are not always visible at quote time. The freight-board form does not show them. They live in the dispatcher's head, if they live anywhere.

5. Everything keeps changing

The matrix is also a moving target. In the last three years: Maut indexed to CO2 classes in late 2023 raising rates 20–40 % for older tractors. Poland's full e-TOLL phase-in in 2021. Bulgaria's TOLL Manager rollout in 2020. Switzerland raised LSVA rates in 2025. France updated péage classifications. Diesel prices across the EU swung 25 % in a 12-month window. Cabotage rules were reinterpreted after the Mobility Package settled.

The mental matrix assembled three years ago is wrong today in five places nobody has checked. Even the colleagues who maintain a rate spreadsheet update it once a quarter at best, and miss new rates between revisions. Maintaining a fully current matrix for 30 countries × every relevant parameter is genuinely full-time work. Nobody who is also dispatching loads is doing it.

Where the math becomes tractable

This is where software earns its place, narrowly: by keeping the per-country price matrix current and applying it accurately to specific route plans. For a planner like Transport Nomad, the cost view shows two lines for any route you plan.

The variable line is fuel (consumption × current per-country price) plus tolls (correct schedule for your axle count, emission class, country segment, at the rate that applies today, not the rate that applied last year). Country boundaries change the rate within a single trip. Ferries and tunnels are at current operator prices.

The fixed-cost line is the trip's share of your fleet's fixed cost, amortised per hour against the trip's actual duration. Driving time comes from the route itself; you can adjust the time up to reflect known loading, unloading or waiting hours that the route alone does not see. The fixed-cost figure recalculates as you change the time, because the underlying number is €/h × hours, not €/km × km.

This is why two 700 km offers with different time profiles produce different cost lines on the same planner — not because of any analytics, but because the math itself is set up to reflect time on the time side and distance on the distance side. The €/km approximation merges them; the planner separates them.

What the planner does not do, said plainly: it does not predict whether the customer will pay on time, whether the delivery will slip 6 hours instead of 2, whether a better load will appear at the destination tomorrow, or whether a competitor will undercut you next month. Those remain judgment calls, and they remain the part of the job where carrier expertise actually lives. The planner removes the calculation step, which is the impossible part at this scale and dynamism, and leaves you with the judgment step, which is where you wanted to spend your two minutes anyway.

The dispatcher's job after this

The in-head calculation is not lazy. It is mathematically impossible at the scale and dynamism of modern European road freight. Thirty countries, multiple parameters each, time-amortised fixed costs, non-linear thresholds, monthly rate changes. The dispatchers who think they "just know" the numbers are running on four corridors of accurate intuition and quietly losing margin everywhere else.

Software exists to turn impossible into thirty-second tractable. The decision still belongs to the dispatcher. The number the decision rests on stops being a guess.

If you want to see what your specific corridors actually cost, with current per-country tolls and your own fleet's fixed cost per hour applied to time, you can try Transport Nomad free for 14 days. The setup is five minutes per vehicle (monthly km + annual fixed cost + working days). After that, every offer you evaluate from the freight board comes with the real number, not the in-head approximation that worked on four corridors and silently failed on the rest.