How Numeo Reads a DAT Load Into Structured Data

A DAT posting is dense with information and almost none of it is structured. Origin and destination sit in one line, the rate in another, equipment as an abbreviation, mileage somewhere off to the side, and the broker name and contact buried in the detail panel. A dispatcher reads all of that, holds it in their head, and re-types the pieces that matter into a spreadsheet or a TMS to compare loads or fire off a quote. Numeo removes that step. It reads the posting where it sits on the board and turns it into clean, typed fields the instant you look at it.

This is the unglamorous work that makes everything downstream possible. You cannot rank loads, flag bad rates, or pull up a broker history if the load is still a wall of text. Numeo Spot, the Chrome extension that runs on top of DAT, parses every posting into the same shape — rate, lane, miles, equipment, broker — so the rest of the dispatch loop has something concrete to work with.

What Numeo pulls from a posting

Every load on DAT carries the same core facts, just laid out for human eyes rather than software. Numeo reads the posting in place and extracts each field into a known slot. The rate comes through as a number, separated from any flat-versus-per-mile ambiguity. The lane resolves into a clean origin and destination with city and state. Mileage is captured as loaded miles. Equipment is normalized from the board's shorthand — reefer, van, flatbed, step deck — into a consistent type. And the broker is identified by company and contact, which is the hook everything else hangs on.

The point is consistency. A human reading two postings back to back will interpret them slightly differently depending on how the broker wrote them. One broker posts "ATL to DAL," another writes out "Atlanta, GA Dallas, TX," a third drops the state. Numeo collapses all of those into the same structured lane, so when you compare ten loads they are genuinely comparable. The same goes for rate: a flat $1,950 on a 620-mile run and a $3.10/mile posting are the same economics expressed two ways, and Numeo reconciles them into one number you can sort on.

Here is what a single raw posting looks like before and after Numeo reads it:

None of this requires you to open the load, copy anything, or switch tabs. The extraction happens on the posting as it renders, so the structured version is ready before you have finished reading the raw one.

Re-keying is the hidden tax

Manual data entry feels like a small cost until you count it across a shift. A dispatcher comparing loads on a tight lane might re-key a dozen postings into a spreadsheet just to see them side by side, and each one is a chance to fat-finger a rate, drop a digit off the mileage, or mix up origin and destination. Those errors are not loud. They surface later as a load that looked better than it was, or a quote sent on the wrong mileage.

The freight workforce is not large enough to waste on transcription. There are roughly 787,000 carriers registered with FMCSA as of December 2023, and the vast majority — 91.5 percent — run ten trucks or fewer, which means the person re-keying loads is often the same person negotiating them, dispatching them, and tracking them. Dispatcher pay sits around $46,860 a year per BLS, and every hour spent retyping board data is an hour not spent on the work that actually moves margin. On a board carrying more than 500,000 loads a day across a network of over 1.7 million trucks, the difference between reading loads and retyping them compounds fast.

Structured extraction kills the tax at the source. Because Numeo reads the field directly off the posting, there is no transcription step to get wrong. The rate you sort on is the rate the broker posted. The mileage you quote against is the mileage on the load. The lane you compare is the lane as the broker intended it, normalized but never re-entered by hand.

Structure is what makes ranking possible

Once a load is in fields, you can do arithmetic on it and that is where the dispatcher actually wins or loses. The single most useful number in spot freight is the all-in rate per mile, and you cannot compute it reliably until rate and miles are both clean. Numeo does that the instant it reads the posting, so a board of loads becomes a sortable list ranked by what each one really pays, not by whatever number the broker chose to display.

That ranking matters because deadhead and posted rate together decide whether a load is worth taking, and the math is unforgiving. Carriers run somewhere between 15 and 30 percent of their miles empty, and a posted rate that looks strong collapses once you factor the deadhead to reach the pickup. With structured fields, Numeo can weigh the all-in cost against a market reference — the ATRI 2025 report put the average marginal cost of operating a truck at about $2.26 per mile for 2024 — and surface the loads that clear it. A dispatcher eyeballing raw postings simply cannot run that comparison across a full board in real time.

The broker field does similar work on the relationship side. Because every load is tied to an identified broker and contact, Numeo can pull that broker's history the moment the load appears: how they have posted before, what they tend to pay on this lane, whether you have worked with them. Instant broker lookup turns a cold posting into a warm one, and it only works because the broker was extracted as structured data rather than left sitting in the detail panel.

How the extracted load feeds the loop

Reading the posting is the first step, not the last. The structured load is the object that flows through the rest of Numeo, and each stage assumes it already exists in clean form.

Filtering and ranking come first. With every load in fields, you can filter the board down to the equipment you run, the lanes you want, and a rate-per-mile floor and have it ranked by all-in economics, instead of scrolling and squinting. Loads with a wide gap between posted rate and market reference float to the top because the structured numbers make that gap visible. This is the difference between hunting through a board and being handed a shortlist.

Negotiation comes next, and Numeo negotiates by email. When you move on a load, the extracted fields — rate, lane, miles, equipment, broker contact — become the basis of the outreach, so the message goes out grounded in the real numbers rather than something retyped from memory. The broker history attached to that load shapes the ask, because you already know how they tend to price the lane. Nothing here is autonomous voice calling; it is structured email negotiation built on data the extension already read off the board.

Finally, the structured load gives every later decision context. When you review what came back, "negotiated to $2.45 on a lane that references near $2.30" means something, because the rate, the miles, and the market figure are all clean fields rather than a guess. The same extraction that saved you the re-keying at the top of the loop is what makes the load legible at the bottom of it.

The takeaway

A DAT board is full of value that is hard to use because it is unstructured. The rate is there, the lane is there, the broker is there — but as text a human has to read, hold, and retype before any of it can be compared or acted on. Numeo's job at the front of the dispatch loop is to read that posting into structured fields automatically, so re-keying disappears, errors with it, and the board becomes something you can sort, rank, and look up against broker history in real time.

Everything Numeo does downstream — filtering by true rate per mile, surfacing underpriced loads, negotiating by email — depends on this first step being clean. Get the load into fields and the rest of the loop has something solid to stand on. See how it works on the board in Numeo Spot.