LTL Cost per Shipment: A Better KPI Framework
Build an auditable LTL cost-per-shipment KPI using final billed cost, eligible shipments, normalized cohorts, and explanatory metrics.
LTL cost per shipment is useful only when the numerator, denominator, and comparison group are defined. A defensible primary KPI is the total included billed cost for a set of completed LTL shipments divided by the number of eligible shipments in that same set. Report it beside shipment-mix, billing, accessorial, and service metrics so a change in the average is not mistaken for a carrier price change.
The framework should answer three questions: What did the business actually pay? What freight and service did that amount cover? What changed from the quote or prior period? A quote amount alone cannot answer them.
Define the KPI before calculating it
Write a short data definition that states:
- which shipments qualify as LTL for this report;
- the origin, destination, customer, facility, or lane scope;
- the shipment-completion event used for inclusion;
- the invoice status required before a shipment is counted;
- the charges included in billed cost;
- how credits, rebills, taxes, duties, claim recoveries, and currency are treated;
- the reporting period and date field used; and
- the freight and service attributes used to create comparable groups.
For an invoice-based transportation KPI, included cost commonly begins with the carrier-billed transportation amount, fuel, and applicable pickup or delivery services. A business may also maintain a broader landed-cost measure that includes internal labor, inventory effects, damage, or other indirect expenses. Keep those as separately named metrics. Mixing them into one unlabeled number makes carrier and period comparisons difficult to audit.
Do not count a shipment merely because it was quoted. Cancelled quotes, duplicate tenders, voided shipments, and unbilled records need their own disposition. If an invoice is still disputed or awaiting a credit, mark the record provisional rather than silently treating the current amount as final.
Use a reconciled shipment record
Create one row per shipment and retain enough information to trace every amount to source evidence.
| Field group | Minimum useful fields |
|---|---|
| Identity | Shipment ID, quote reference, bill of lading reference, invoice reference |
| Lane | Origin, destination, facility types, pickup and delivery dates |
| Freight | Handling units, packaged dimensions, gross weight, commodity, NMFC item, class, stackability |
| Service | Carrier, service level, required accessorials, purchased guarantee if any |
| Quote | Submitted freight facts, quoted transportation amount, fuel, services, effective date |
| Invoice | Billed transportation amount, fuel, accessorials, adjustments, credits, final total |
| Outcome | Delivery result, damage or shortage notation, claim, billing dispute and resolution |
Reconcile the quote, freight as tendered, and invoice. A billed difference can result from changed dimensions, weight, class, location type, services, dates, or the application of a governing carrier rule. It can also be a correctable billing issue. Classify the cause instead of assuming every variance is carrier error.
ArcBest’s current shipping-data guidance identifies costs, delivery times, routes, freight size and weight, damage, and late deliveries as useful operational inputs. That is a practical starting set, but the business still needs its own definitions and source records.
Make billed all-in cost per shipment the anchor
For a defined cohort:
Billed all-in cost per shipment = total included final billed cost / eligible completed shipments
Use the same inclusion rules for both parts of the formula. If a credit belongs to an included shipment, assign it according to the documented accounting rule. If a shipment is excluded because it is not final, exclude its current cost as well.
Show the shipment count with the average. Also show the median and range when volume permits, because one unusual shipment can move the mean. Never present a change in cost per shipment without describing changes in weight, dimensions, handling units, class, lane, services, or carrier mix.
The KPI is best calculated for comparable groups such as the same lane and facility types, similar packaged freight, the same service requirements, and a consistent pricing context. A company-wide number can be a financial summary, but it is rarely enough to diagnose an LTL operating change.
Add companion metrics that explain the average
Quote-to-bill variance
Measure the final billed total against the accepted quoted scope, then assign a reason code. Report both the amount and frequency of variance. Keep shipment-data changes, added services, carrier-rule applications, duplicate charges, and corrected bills separate.
The metric should not imply that a quote was unconditional. Compare the submitted and tendered freight facts before labeling a difference inaccurate.
Accessorial incidence and cost
For each service type, report:
- shipments using the service;
- share of eligible shipments using it;
- included billed amount; and
- quote-to-bill variance attributable to it.
Accessorial definitions and charges are carrier-specific. ArcBest’s current LTL pricing guide identifies services such as residential, liftgate, and inside delivery as additions beyond a standard shipment. Use the rules and pricing agreement that actually govern each shipment.
Cost per hundredweight, handling unit, or pound
These can be useful normalization metrics inside a comparable cohort:
- billed cost divided by hundredweight;
- billed cost divided by handling units; or
- billed cost divided by gross pounds.
None is a universal carrier-performance score. A light high-class shipment, a heavy dense shipment, and a non-stackable shipment can produce different ratios for valid reasons. Display commodity, class, cube, weight, lane, and services with the metric.
Shipment-mix indicators
Track the variables that can move the average even when pricing is unchanged:
- gross weight and packaged cube;
- handling units;
- freight class and commodity mix;
- stackable versus non-stackable;
- lane and facility-type mix;
- accessorial frequency;
- service-level mix; and
- carrier and pricing-program mix.
Current NMFTA guidance describes freight class through density, handling, stowability, and liability. That is why weight alone cannot normalize shipments with different classification treatment.
Service and exception outcomes
Cost needs operational context. Depending on the business need, track:
- pickup completion against the requested window;
- actual transit time;
- delivery performance against the applicable estimate or purchased commitment;
- damage and shortage notation;
- claim frequency and disposition;
- billing disputes and correction time; and
- tracking-event completeness.
Estimated transit is different from guaranteed service. The KPI definition should record what was purchased and which carrier terms applied before a shipment is classified as on time or late.
Separate carrier price from shipment mix
When cost per shipment rises, test the explanations in order:
- Did average weight, cube, handling-unit count, commodity, or class change?
- Did the lane, destination type, accessorial, or service mix change?
- Did carrier allocation or the pricing model change?
- Did fuel, minimum charges, tariffs, or carrier-specific rules change?
- Did quote-to-bill adjustments or unresolved disputes increase?
- Did the underlying comparable carrier price change after those effects were controlled?
Carrier tariff structures demonstrate why the distinction matters. FedEx Freight’s current 100-Y Rules Tariff contains carrier-specific standard and special-service charges, classification exceptions, and pricing-program provisions. Old Dominion’s current 6100-B discount tariff applies discounts under stated governing publications and generally excludes accessorial charges from those discounts. These are examples, not rules for the whole market. Apply only the documents and agreement governing the shipment being analyzed.
A scorecard that stays interpretable
Use a compact scorecard for each comparable cohort:
| KPI | Result to show | Context beside it |
|---|---|---|
| Billed cost per shipment | Mean, median, range, shipment count | Lane, period, freight and service scope |
| Quote-to-bill variance | Amount and frequency | Reason-code breakdown |
| Accessorials | Incidence and billed amount by type | Facility and service mix |
| Normalized cost | Per hundredweight, pound, or handling unit when useful | Weight, cube, class, commodity |
| Service | Outcome against defined estimate or commitment | Carrier and service level |
| Exceptions | Damage, shortage, claim, dispute | Disposition and unresolved count |
Avoid a single blended score that hides trade-offs unless every weight and threshold is documented. A decision maker should be able to see whether a lower cost came with different freight, fewer services, or a different service outcome.
A worked comparison without invented prices
Suppose a business compares two periods on one recurring lane. The reported cost per shipment is higher in the second period.
The shipment table shows that the second period contained more handling units per shipment and more destination appointments. After the analyst separates appointment shipments and compares similar handling-unit groups, the billed transportation amount for the core group is broadly consistent. Several quote-to-bill variances also trace to outdated packaged dimensions in the quoting record.
The correct conclusion is not that the carrier raised the lane price by the same amount as the top-line KPI. The business should correct its shipment data, retain the appointment cost as a real service requirement, and compare current like-for-like quotes. The KPI detected a change; the companion metrics explained it.
This example provides a method, not a benchmark, savings estimate, or claim about a specific carrier.
Handle edge cases explicitly
- Credits and rebills: attach them to the original shipment or follow one documented period rule.
- Open disputes: mark the shipment provisional and report the unresolved amount separately.
- Claims: do not net a claim recovery against freight cost unless the metric definition explicitly requires it.
- Cancelled or failed pickups: keep operational costs visible, but do not count them as completed shipments without a named metric.
- Multi-currency data: use a documented conversion source and date; retain the original currency and amount.
- Mixed modes: keep parcel, volume LTL, partial truckload, full truckload, and expedite out of an LTL cohort unless the report is intentionally multimodal.
- Sparse cohorts: show individual results or widen the period rather than presenting an unstable average as a trend.
Prepare a shipment-specific quote comparison
Bring the final origin and destination, facility types, pickup date, handling-unit count, packaged dimensions and gross weights, commodity descriptions, current NMFC items and classes when known, truthful stackability, and all required pickup or delivery services.
Shipocity is backed by a team with more than 40 years of combined logistics experience. Through established industry relationships, the platform helps businesses compare competitive freight rates for their specific shipment.
With one verified shipment record, compare live carrier rates.
