There are three ways to get a number for an LTL shipment. Most operations pick one and live with its failure mode without ever naming it.
Direct carrier APIs
You hold credentials with each carrier and call them for a rate.
Good at: accuracy. It’s the carrier’s own answer, using their current tariff, their fuel index and your negotiated pricing. Nobody is between you and the number.
Breaks when: the API is down, the API is incomplete, or you don’t have a relationship. Outages happen and quoting stops. Incompleteness is worse because it’s silent — published accessorial lists that don’t match what actually prices, codes that exist in the tariff but not the documentation. And every new carrier is a separate integration, separate credentials, separate quirks.
Rate consolidators
One integration gives you many carriers.
Good at: coverage and speed to launch. One connection, dozens of carriers, someone else maintaining them. For a broker starting out this is the difference between quoting this month and quoting next year.
Breaks when: you need something the consolidator’s model doesn’t carry. They normalize many carriers into one shape, and normalization loses detail — carrier-specific accessorial logic, unusual rules, the rate quirks that only matter on the loads where they matter. You’re also dependent on their coverage decisions and their uptime, and there’s usually a spread.
Internal rating logic
You load a base tariff — SMC3 or similar — plus your discounts, fuel table, minimums and FAK exceptions, and calculate the rate yourself.
Good at: it always works. No API to be down, no third party. It’s also the only method that can price a shipment nobody will quote you on, and the only one where you can see exactly why the number is what it is, line by line.
Breaks when: your inputs drift. A discount tier resets, fuel changes, an FAK exception expires, the carrier issues a GRI. The engine keeps producing confident numbers off stale inputs, and nothing tells you.
The pattern in those three
Each is strong where the others are weak. Carrier APIs are accurate and fragile. Consolidators are broad and lossy. Internal logic is available and drifts.
The question isn’t which one is best. It’s what happens when the one you’re using is wrong — and with a single source, the answer is always the same: nothing happens. You get a number, it looks like every other number, and you find out when the invoice arrives.
Why running all three changes the problem
Rate the same shipment three ways and disagreement becomes information.
If the carrier API and your internal calculation agree, confidence is high. If they disagree by twelve percent, something is wrong right now, before the load moves — a discount that reset, an accessorial the API applied that your tariff doesn’t know about, a fuel base mismatch. That’s a signal you can act on rather than a discrepancy you discover later.
It also means an outage doesn’t stop you quoting. Internal logic carries the load, flagged as a fallback rather than presented as a carrier-confirmed rate.
And it catches unapproved increases. A carrier that pushes a GRI you didn’t agree to shows up as a growing gap between their API and your contracted calculation. With one source, that increase is invisible — it just becomes the new price.
What “reconciled” has to mean
Running three sources is only useful if something reconciles them. Three numbers in three tabs is not a rating strategy.
Reconciliation means a rule set: which source wins by default, what size of disagreement triggers a flag, what gets shown to the person quoting, and what gets logged for later. Without that you’ve added complexity and no decision support.
It also means the same logic that produced the quote is available afterwards to check the invoice against. If quoting and auditing run on different engines, every comparison is manual, and manual comparisons happen on the loads somebody remembers to check.
The maintenance nobody budgets for
Whichever you pick, something needs looking after, and the cost is not in the integration — it is in keeping it true.
Carrier APIs need credential rotation, and they change response formats without much notice. Consolidators need their coverage checked against the carriers you actually use rather than the ones on the marketing page. Internal rating needs the base tariff updated, discounts refreshed at renewal, fuel tables current and FAK exceptions reviewed before they expire.
The failure in all three cases is silent. Nothing errors. You just start quoting off something slightly wrong, and you find out from an invoice. A quarterly check that re-rates a sample of historical shipments and compares the result to what was actually billed catches most of this in an afternoon.
What this costs
Honestly: more than one integration.
Direct carrier connections take credentialing and maintenance. Internal rating means owning a base tariff and keeping discounts, fuel and exceptions current. That’s real work, and for a small book it may not pay.
The threshold is roughly where a single bad quote costs more than the maintenance. If you’re quoting a handful of LTL loads a month, a consolidator alone is fine. Once LTL is a real line of business, single-source rating stops being simplicity and starts being an unmonitored risk.
The short version: three methods, three failure modes, and the one you chose is invisible until it’s expensive. Running them together turns a silent failure into a flag.
That’s what the hybrid rating engine does — carrier APIs, your own loaded contracts, and internal logic on a base tariff, reconciled into one quote. The mechanics of how it rates are here, and the same engine checks the invoice against what it quoted.
So here’s the question worth sitting with: if your rate source were wrong by ten percent today, what would tell you?