The Cheapest Quote Should Not Choose the Carrier
Key Takeaways
- The cheapest quote should not choose the carrier. Its position on two axes should: how its full expected cost sits against the lane target and approval limit, and whether it meets the shipment's hard service requirements.
- Crossing 3 cost positions with 3 service positions gives 9 cells, and each cell maps to one fixed outcome: book, negotiate, clarify, route to an approver or reject.
- A quote that fails a hard service requirement is rejected in every cost position. Price never rescues it; only a person who changes the requirement can.
- Wilson, by Cartage, is an AI logistics agent that requests quotes from approved carriers, negotiates against the company's targets and books within the company's approval limits, escalating anything outside them.
- Wilson keeps no standing carrier scorecard. It can produce carrier-performance reports on request from the loads it runs, which helps a team review where its carriers keep landing in the matrix.
What should choose the carrier, if not the cheapest quote?
The cell each quote lands in should choose the carrier. Place every quote on a cost axis and a service axis, and let the cell decide whether the load is booked, negotiated, clarified, sent for approval or rejected.
Most carrier selection breaks down in the same spot. The team knows price is not everything, but the moment a load needs a carrier, the trade-off between price and service is made from scratch by whoever is on shift.
A matrix replaces that judgment with a fixed outcome per situation. It does not rank carriers or weight scores. It asks two questions about each quote among the company's approved carriers, then reads the answer off a grid that the logistics lead and finance agreed on before the quotes arrived. The value is in the outcomes: each cell already says who acts, and whether a person needs to see the load at all.
Why price alone loses money is covered in the lowest carrier quote shouldn't always win, and which carriers may quote at all is covered in the approved carrier selection rules framework. This grid is the step after both: it takes the eligible quotes and turns them into a booking outcome.
The cost × service × approval matrix
The cost axis has 3 positions, measured on full expected cost (the quote plus the accessorials this shipment will need):
- At or under target: at or below the lane's rate target.
- Within the approval band: above target, but at or under the approval limit.
- Above the limit: above the approval limit.
The service axis also has 3 positions:
- Meets: the carrier has confirmed every hard requirement (equipment, pickup date and window, delivery date, appointment, required accessorials).
- Open item: nothing fails, but at least one requirement is unconfirmed, such as an appointment not acknowledged or a liftgate not mentioned in the quote.
- Fails: the carrier cannot meet a hard requirement.
| Service position | At or under target | Within approval band | Above the limit |
|---|---|---|---|
| Meets | Book | Negotiate, then book if the counter stays inside the band | Negotiate; if still above, route to approver |
| Open item | Clarify, then book | Clarify and negotiate, then book | Route to approver with the open item stated |
| Fails | Reject | Reject | Reject |
When several quotes land in cells that allow booking, the one with the lowest full expected cost wins. The bottom row is the reason the matrix exists: a quote that cannot meet the pickup date is not a cheaper option, whatever its price.
What has to be set before the matrix can run
The grid only works if its edges are written down per lane before any quote comes in.
- Rate target and approval limit. Each lane needs both numbers. The band between them is where negotiation happens without a person; above the limit, a named approver decides. A band that is too narrow sends routine loads to approval, and one that is too wide removes the reason for having a limit.
- Hard requirements per shipment. These come from the order and the consignee: pickup date, delivery date, appointment needs, equipment, accessorials. Requirements that are soft preferences should not sit on this axis, or every quote ends up in the open-item row.
- Accessorials folded into cost. If the consignee always needs an appointment, that charge is added to every quote before it is placed on the cost axis. Otherwise a quote that omits it lands one column too far to the left.
- Named approvers per cell. The approver for an above-limit quote that meets service may be the logistics lead. The approver for an above-limit quote with an open item may need finance as well. Write it into the cell.
- A clarification deadline. Open items need a time limit tied to the pickup date. A quote still unconfirmed at the deadline moves down to the fails row.
- A quote cutoff. Without one, the matrix waits on a carrier that may never answer.
Decision rules for the edge cases
The 9 cells cover most loads. A few situations need a written rule of their own.
- No quote reaches a booking cell. If every quote is in the fails row or above the limit after negotiation, the load goes to a person with all quotes and their cells shown. The options are a different date, a higher approved rate, another mode or a newly approved carrier. Each is a business decision. For what happens when nobody accepts at all, see what happens when no approved carrier accepts the load.
- The preferred carrier is silent at the cutoff. Silence is not a cell. The matrix runs on the quotes that arrived. A late reply can still be placed if the load is not yet booked.
- The only quote fails service. It is still rejected by the matrix. A person can relax the requirement, for example by moving the delivery date with the customer, and the quote is then placed again.
- The lane target is stale. If most quotes over several weeks land above the limit, the target needs review, not repeated approvals. A lane-rate report on the loads already run is the input for that review.
- Urgent loads. Urgency should move the approval limit for that load, not the service axis. A rushed load still has to make the pickup.
Worked example: four quotes, four outcomes
The figures are illustrative. A plastics manufacturer needs an inbound FTL load of resin for a production run. Pickup must happen Tuesday, and the plant requires a delivery appointment. The lane target is $2,400 and the approval limit is $2,600.
| Carrier | Full expected cost | Service position | Cell outcome |
|---|---|---|---|
| D | $2,150 | Fails: pickup Wednesday at the earliest | Reject |
| E | $2,380 | Open item: appointment not acknowledged | Clarify, then book |
| F | $2,520 | Meets | Negotiate, then book inside the band |
| G | $2,710 | Meets | Negotiate; route to approver if still above $2,600 |
D is the cheapest quote and leaves the matrix first, because a Wednesday pickup stops the production run. E is the best candidate, so the appointment question goes to the carrier with a deadline. If E confirms, the load books at $2,380 with no approval needed. If E does not confirm by the deadline, E moves to the fails row, and F becomes the candidate: a counter toward $2,400 follows, and F books at any agreed rate up to $2,600 without approval. G only reaches an approver if both E and F drop out and G will not come down. In that case the approver sees all four quotes, their cells and what was already tried.
How Wilson, by Cartage applies the matrix
Wilson, by Cartage, executes the matrix on each load: it gathers the quotes, places them against the company's targets and rules, negotiates inside the band, books what the rules allow and sends the rest to the named person.
- Quotes: Wilson requests quotes only from the company's approved carriers and follows up on missing replies. Cartage can introduce additional capacity, but the company approves any carrier before Wilson uses it.
- Negotiation: Wilson negotiates against the company's rate targets under configured rules and approval limits. Rates that stay outside those limits go to the logistics team.
- Booking and approvals: Wilson starts with assisted execution, and the company configures which outcomes need approval. Routine bookings inside the rules can run automatically as the team gains confidence in the workflow.
- Escalation: if no approved carrier accepts, Wilson escalates with the quotes and steps already taken.
- Record: quotes, messages, confirmations and changes are recorded in the shipment's change history, so the reason a carrier won can be checked later.
Wilson keeps no standing carrier scorecard. Teams that want to see which carriers keep landing in the open-item row can ask for a carrier-performance report from the loads Wilson runs.
To test the grid, bring three real quotes and one hard requirement to Cartage and see which cell each one lands in.
Sources
- Cartage, Wilson product FAQ and product scope, October 2026: https://cartage.ai