Picture a forwarder’s quote that lands in your inbox: $10,147 per 40-foot container, Ningbo to Dallas, all-in. One number. Sounds reasonable. Maybe it is, maybe it isn’t, but you can’t tell from the number alone.
A should-cost model decomposes that one number into the 15+ cost drivers stacked underneath it. The base ocean freight, the bunker surcharge, the origin THC, the destination drayage, the chassis rental, the customs brokerage, the duty, the MPF + HMF, the expected D&D. Each one has its own driver, its own benchmark, its own negotiability.
The goal isn’t to replicate the carrier’s pricing model. The goal is to know which levers move the number, which don’t, and where the quote is strong or weak.
Starting cold? “Should cost” is the practice of building up a quote from first principles so you can pressure-test the carrier’s number, identify where they’re padding, and know which levers you can pull.
Inbound freight is where most should-cost practices are weakest, because:
The deep-dives that follow apply the same structure to each cost bucket: cost build-up → sensitivities → where to benchmark → shipper levers → hidden-cost map → worked example. Origin logistics and customs brokerage are treated as their own buckets alongside the main-carriage modes, because they’re separately-invoiced, separately-negotiated costs that most quotes under-specify.
Every mode-level should-cost page here follows the same six sections. Use them as a template when you build one for a lane you care about.
The line-by-line decomposition. Each row names the driver, a typical range, whether it is fixed / indexed / spot / negotiable, and where to benchmark it.
The 2–4 drivers that account for most of the variance on a real lane. These are where negotiation effort and operational change produce results.
Named public indices, carrier tariffs, and subscription data sources. Most mode-level rate transparency is concentrated in 3–5 sources per mode.
What the shipper can actually change, ranked by typical impact. Distinguishes contract-level levers (RFP, carrier mix) from operational levers (dock scheduling, classification, packaging).
The line items a “competitive” quote typically strips out. These often account for 10–30% of real landed cost and are where the cheapest-looking quote turns out not to be.
A concrete lane, end-to-end, with a visual cost waterfall showing the composition. Hard numbers are more useful than percentages once you’re in a negotiation.
Imagine you’ve built a clean should-cost model and an annual RFP comes in with five carrier responses. How do you actually use the model? Four practical patterns:
Force every bidder to price the same scope, line-item, with explicit exclusions. Bid responses that quote different scopes can’t be compared directly, and that’s often where bidders hide margin. A should-cost build is the leveling template.
Plot every bidder’s response against your should-cost ranges. The interesting bids are the ones that fall outside the range, in either direction.
Either way, the outlier bid is a question, not a verdict. Open the conversation; don’t just accept or reject.
The incumbent has information advantages: lane history, your dock behavior, your accessorial profile, your operational quirks. They can price tightly. A challenger is pricing partly on hope.
A standard pattern: incumbent bids ~5–10% above their actual cost; challenger bids ~5–10% below their actual cost; you compare them and the incumbent wins on price by 2–5%. Without a should-cost model, that’s a clean win for the incumbent. With one, you can see that the challenger is pricing below sustainable cost (red flag for service quality at year 2) or that the incumbent has padding worth re-negotiating regardless of who wins the lane.
For modes with public indices (ocean FBX/Drewry, road DAT, air TAC), index-relative benchmarking on your should-cost build catches drift over a contract period. “Our incumbent quoted FBX+15% in 2024 and is pricing FBX+22% in 2025 with no scope change” is a defensible negotiation position.
Cheapest-looking bid is rarely the cheapest in practice once D&D, accessorial profile, and reliability-driven safety stock are counted. Your operations team has gut-feel data on which carriers run clean lanes and which don’t. Cross-reference operator intelligence with should-cost output before tier-ranking bidders. The data plus the operator intel is more defensible than either alone.
Origin logistics
The factory-dock to port-gate stretch: origin drayage, CFS or buyer’s-consolidator receiving, origin VAS, palletization, stuffing supervision. Typically 3–10× the quoted “origin drayage” line on an ocean bill and living on a separate invoice.
Customs brokerage
The service category that spans every mode. Tier-1 vs. Tier-2 vs. captive brokerage decisions, HS classification audit value, post-entry support, and where the brokerage line on your quote actually understates real cost.
Ocean FCL
Full container load on intercontinental lanes. Base freight is ~18% of total landed cost; the rest is what the should-cost build exposes.
Air
General cargo airport-to-airport-plus-handling. Chargeable weight and packaging density dominate the cost equation.
Road — FTL & LTL
US truckload and less-than-truckload. The mode with the deepest public benchmarks (DAT, Cass, Sonar) and the most execution leverage (dock discipline, classification).
Rail & intermodal
Domestic and international intermodal. True cost is dominated by the drayage bookends and rail reliability’s effect on safety stock.
Parcel & express
The most opaque mode; effective cost is sensitive to dim divisors, zone, surcharge stack, and the ~40 published but rarely-negotiated fees.
Specialty modes
Breakbulk, ro-ro, project cargo, pipeline. Where “no public benchmark exists” is the defining fact, and scope-of-work discipline matters more than rate.