In short: A contracted freight rate is an option the carrier holds, so it survives while capacity is loose and gets declined once the lane moves against them, which is why the saving booked at award is not the saving realised over the year. Splitting a lane between a primary and a backup costs a little on the blended rate and absorbs the fallout at an awarded rate instead of at spot, and it can return more than the cheaper single award did. Caplice and Sheffi set out the optimisation-based bid format in the Journal of Business Logistics in 2003, where carriers price bundles conditionally and the buyer solves for the whole award set. Routing guide compliance, the share of loads that moved on the primary awarded carrier at the awarded rate, is the test of whether a rate structure describes what you paid.
The annual tender comes back with rates eight percent below the incumbent. Procurement books the saving, the contract is awarded, and for four months everything is fine.
Then the market tightens. The carrier who bid aggressively to win volume starts declining loads on the lanes where their rate is now below market. Your tender rate is technically still in force and it is unenforceable, because the remedy for a declined load is to find another truck at spot, and spot is above what the incumbent was charging before the tender.
The eight percent was real on the day it was signed. Whether it was real over the year depends on things the tender did not measure.
What a rate actually is
A contracted freight rate is a price at which a carrier has said they are willing to move a load, under conditions neither side fully specified.
The conditions matter more than the number. A carrier prices a lane based on what they can do with the truck afterwards, which is why a lane into a region with plenty of outbound freight prices lower than the same distance into a region with none. It also depends on how predictable your volume is, how long they wait at your dock, and whether your loads fit their existing network.
Two consequences follow.
A rate quoted for a lane you tender occasionally is not the same product as a rate for a lane with steady weekly volume, even at the same origin and destination. The carrier priced the second one assuming they can plan around it.
And a rate accepted at a level the carrier cannot sustain will be honoured while capacity is loose and quietly abandoned when it is not. Award coverage that looks cheap on paper frequently carries an implicit option the carrier holds and you do not.
This is why transportation procurement research moved toward bid formats that let carriers price bundles instead of isolated lanes. Caplice and Sheffi set out the optimisation-based approach in the Journal of Business Logistics in 2003, where carriers submit conditional bids expressing what a lane is worth given what else they win, and the buyer solves for the whole award set rather than picking a winner lane by lane. Most tenders still ask for a rate per lane in a spreadsheet, which forces every carrier to guess at the network they will end up with and to price that guess conservatively.
Designing the tender
Three design decisions do most of the work, and all three are usually made by default.
Lane granularity. Tendering at a postcode-to-postcode level produces thousands of lanes, most with too little volume to be priced properly, and carriers respond with padded rates on the thin ones. Aggregating to regional lanes gets better pricing and less precision. The sensible compromise tenders the top lanes individually and groups the tail, which requires knowing where the volume concentration sits before the tender opens. Finding that is a two-hour exercise: sort last year's loads by lane, take the cumulative share, and look for where the curve flattens. In most networks a few hundred lanes carry the bulk of the volume and several thousand carry the remainder, and those two populations want different treatment in the bid document.
Volume commitment. A tender with no committed volume gets priced as an option. A tender with a genuine commitment gets a better rate and constrains you. Most businesses commit to nothing and then complain that carriers do not honour rates, which is a reasonable outcome from an unreasonable ask. A commitment that means something carries a floor, a look-back period and a stated consequence for missing it. Language along the lines of estimated annual volume subject to business needs reads to a carrier as an option you have asked them to write for free, and they price it with a margin they keep whether or not the volume appears.
Award structure. Awarding one hundred percent of a lane to the cheapest bidder maximises the paper saving and removes your fallback. Splitting a lane across a primary and a backup costs a little on the blended rate and buys you a carrier who is already set up when the primary declines. On volatile lanes that insurance is usually worth more than the spread.
Put arithmetic on that last one. A lane runs 5,000 loads a year. The incumbent charges 1,850 and the winning bid is 1,702, eight percent lower, so the tender books a saving of 740,000.
The first half of the year runs as awarded. In the second half the market tightens and acceptance on that lane falls to 72 percent, so 700 of the 2,500 loads go to spot at 2,220. The year costs 2,500 times 1,702, plus 1,800 times 1,702, plus 700 times 2,220, or 8.87 million against 9.25 million under the incumbent. The saving was real and it came in at 4.1 percent against the 8 that went into the tender report, and nobody went back to correct the report.
Now award the same lane 70 percent to the low bidder and 30 percent to a backup at 1,790. The blended rate is 1,728.40, one and a half percent worse on paper, which is enough to lose the tender presentation. When the market tightens the backup absorbs the fallout at its awarded rate rather than at spot, so the year costs 8.64 million and the realised saving is 6.6 percent. The award structure that looked more expensive returned more than half again as much.
Benchmarking without fooling yourself
Rate benchmarks are widely available and easy to misuse. Two things to check before drawing a conclusion.
Comparable scope. A benchmark rate is usually line haul, and your invoice includes fuel, accessorials, detention and possibly a fixed capacity commitment. Comparing your all-in cost against a line haul benchmark makes you look expensive by a margin that is entirely definitional.
The same definitional gap hides inside the tender itself. Two carriers quote an identical line haul. One escalates fuel from a 1.25 dollar diesel base at 6.5 miles per gallon, the other from 1.10 at 6.0. At four dollar diesel that is 42.3 cents a mile against 48.3, which on a 600 mile lane is 253.85 against 290.00. Thirty-six dollars a load, invisible in a bid comparison that only lists line haul, and 180,000 dollars across 5,000 loads. The symptom is a lane whose cost per mile drifts up while the awarded rate has not moved, and the fix is to normalise every bid onto one fuel table before comparing anything.
Comparable conditions. A benchmark drawn from a market where drop trailers are standard is not comparable to an operation where drivers wait to be live-loaded. If your average detention is ninety minutes, you are buying a different service than the benchmark describes, and the gap is your dock rather than your procurement.
The more useful internal benchmark is your own rate history by lane against a published market index, looking at the spread rather than the level. A lane whose spread to market has widened over two years is one where you have stopped being competitive without anyone noticing, and that comparison is available from data you already hold.
The contract and spot mix
Most freight strategies land somewhere between fully contracted and fully spot, and the right point moves with the market.
Contracted rates are more stable and are usually above spot in a loose market and below it in a tight one. That relationship is the whole argument. Contracting is insurance against a tight market, paid for by giving up the saving available in a loose one.
The practical approach treats it as a portfolio. Contract the base volume you are confident about, on the lanes where a capacity failure would be most damaging, and leave a deliberate share to spot. The share is a risk decision rather than a cost minimisation, and it should be set with a view on where the market is heading rather than on where it has been.
A rough rule for setting the share beats intuition. Identify the lanes where a capacity failure stops a line or misses a customer commitment, contract those up to the volume you are confident of at the trough of your seasonality, and let everything above the trough run on the market. That puts the contracted book where it is genuinely buying insurance and leaves the discretionary volume where the loose-market saving actually lives.
One measurement makes this decision much easier and is rarely tracked: your actual routing guide compliance, meaning the share of loads that moved on the primary awarded carrier at the awarded rate. A guide with ninety-five percent compliance is a real cost base. One running at sixty percent is a paper rate structure with an undisclosed spot programme underneath it, and the reported tender saving is fiction.
What to measure beyond rate
Four things that determine total cost and rarely appear in a tender scorecard.
Tender acceptance, by carrier and by lane, tracked over time. This is the leading indicator of a carrier who is about to become unavailable, and the slope matters more than the level. A lane holding steady at 91 percent is stable. One that has slid from 96 to 88 over six weeks is a carrier working out that the rate no longer covers the lane, and the weeks between noticing that and the first refused load in peak are when you still have options.
On-time performance against the appointment, measured as a distribution rather than an average, since the tail is what causes downstream cost.
Claims and damage rate, which for some product categories exceeds the rate difference between carriers.
Invoice accuracy, because a carrier with a low rate and a high dispute rate consumes finance time that nobody costs.
A carrier scorecard combining these with rate gives a different ranking than rate alone, and the ranking is more stable year to year, which is itself informative.
Where this stops
Freight procurement operates on a cost base largely determined before the tender opens. Your lane structure, your order profile, your dock efficiency and your lead times set the price band that any carrier can quote inside, and a tender optimises within that band rather than moving it.
If your rates are persistently above market and the tender process is competently run, the answer is usually in the operation rather than in the negotiation. Loads that arrive as partial trailers, appointments that are missed, docks that hold drivers for two hours, and lanes that run heavy in one direction and empty in the other all price into the rate, and all of them are yours to change.
There is also a limit on how far cost reduction should be pushed with a carrier base you depend on. A carrier operating on margins that do not support fleet renewal is a service failure being deferred, and the failure lands at the point you can least afford it. That is not an argument against negotiating, and it is an argument for knowing roughly where the floor is.
Start by measuring routing guide compliance for the last quarter. If it is below eighty percent, your contracted rate structure is describing something other than what you actually paid.