The hardest problem I worked on at DAT

Ownership Before Form Factor

John Leo Szrejter · Perspective

How I reasoned about the next generation of American road freight matching, and why the obvious move (build the unifying product first) is the wrong place to start.

Everyone in American road freight is jumping between five SaaS systems to move a single load. Shippers, brokers, and carriers all burn time stitching together tools that were never built to talk to each other. The obvious fix is to build the one platform that finally unifies them.

That instinct, build the unifying product, is where most platform strategy starts. It is also where most of it goes wrong, because it treats form factor as the first decision. Form factor is the last decision.

Here is the thesis: what you can own determines what form factor can win. Before you design anything, you map who already owns each high-trust decision point across the freight journey, then decide what to own, what to partner for, and what to buy. The product that wins is the one your ownership position actually lets you build. Maturity, meaning how far you can reliably push automation at a given point, gates the rest. Start with the product you wish you had and you will design something your ownership position cannot support.

Where I was sitting

I led research for DAT One, the load board behind 65 to 70 percent of the U.S. spot freight market, a $350 to $400 million annual-revenue product running near 40 percent EBITDA with ties reaching into contract freight. I owned its research landscape and was tasked with identifying the strategy for the next generation of American road freight using everything we had: an internal research team, a $5 million Roper-funded research engagement with Bain, and more first-party freight data than anyone else in the market.

None of that was required to see the environment was fragmented. It was required to decide what to do about it.

The method: map who owns each trust point

I built an end-to-end experience map of the entire freight journey: Shipper Prospecting, Mutual Capacity Matching, Freight In Route, and Freight Arrival at Destination. At every stage I mapped what we owned and led, what we did not, who the personas were, who the incumbent leaders were, the current quality of the experience, and how hard it would be to move a user from one product to the next influenced by what we knew so far. My product partners mapped the other half of the board: each leader's market penetration, the difficulty of transitions between products, and the profit-to-team-cost ratio of each play.

The decision points are not equal. Some carry far more trust weight than others, which means whoever owns that point sets the standard everyone downstream has to follow. Tracking is the clearest example. The party who chooses the tracking tool effectively decides it for the others. Mapping where trust concentrates is what tells you which points are worth owning and which you can safely partner around.

From map to ownership strategy

The question every C-suite asks in a blue-sky landscape is "do we build or buy this?" The real answer in freight is usually "partner, until you can't." You partner until the friction of depending on someone else outweighs the cost of owning it. Transition abrasiveness, meaning how painful it is to move users or data across a seam, is what usually triggers the first serious build-or-buy argument. Once you decide to own, market penetration, product health, and EBITDA over time tell you which company to buy rather than build. And if your long game is training AI, you want to own the surfaces you intend to train on, because you cannot reliably mature automation on a surface you do not control.

Initial Triangulation

Three signals fed the map: my internal research, the $5 million Roper-funded Bain engagement, and my own field signal. I weighted them by confidence, not by cost. My internal work carried the highest internal and external validity because our research rigor had outrun previous teams'. The Bain engagement was expensive but came with a catch: Roper funded it without buying rights to the underlying data, so we could not inspect how the vendor ran its research or fully audit its findings. Historical research and vendor findings were treated as low-to-medium confidence and directionally useful. My recommendations were treated as medium-to-high confidence and usually adopted as the default read. Where we owned the surface, I used internal qualitative and quantitative signal reduced to high, medium, or low confidence. Where we did not, I filled the gap with work-shadow observation, interviews, surveys, and public competitive analysis.

The TruckerTools decision

This is how the framework picks a path.

My research influenced deprioritizing DAT's own tracking software in 2023 and set up the acquisition of TruckerTools. Three factors drove it. First, across studies with shippers, brokers, and carriers, shippers had the loudest voice in choosing the tracking tool and brokers had the second; tracking was a high-trust point owned from the top down, not something we could win by building a better in-house feature. Second, TruckerTools already had a robust tracking product, deep adoption across other product lines, and a large carrier base with contract-freight ties that DAT lacked. Third, our cloud did not need to house tracking itself. It could integrate with it, which made partnering or acquiring clean from a technical-dependency standpoint. A high-trust point we did not own, an incumbent who already owned it well, and a low integration cost. Buy, do not build.

I ran every build-or-buy call through the same reusable template: four dimensions, scored per option, on a common scale. Run the tracking decision through it and Buy carries the board, not just the conversation.

TruckerTools was the first node DAT chose to own on this logic, and the same template shaped what came next.

From raw research to the map

Competitive wargaming

With the map and matrices built, I ran the right-to-win exercise with product and C-level leadership. We started from the experience map, the matrices, and the user flows, walked them together, wrote out the highest-confidence scenarios, and attacked each one for weakness until a single scenario held up.

To grade each build-or-buy scenario consistently, we scored it against a four-dimension evaluation matrix:

  1. Speed and time-to-market. How fast the play could realistically reach customers.

  2. Control and customization. How much we could shape the surface to our needs versus inheriting someone else's roadmap.

  3. Total cost of ownership. The full cost over time, not just the acquisition or build price.

  4. Risk management. Technical, market, and dependency risk across the life of the decision.

The framework was bottom-up before it was top-down. Historical learnings and field signal funneled into it first; UXR and product started the matrix and populated it with evidence, and executive leadership then led the scoring and final call. That sequence mattered: the people closest to the workflow built the case, and the people accountable for the bet made the decision on a common scale rather than on instinct.

From there we listed the jobs to be done, what an MVP had to have, and the nice-to-haves we already knew about. That package went to C-level for prioritization: each leader scored every effort one to five, and the ranked output came back to us for planning.

Sizing the market, by broker segment (2024)

Before any build or buy call, there was a prior question: which markets to aim at, and how hard. I built the answer with the senior director of product as a director-level market-prioritization tool, not a feature spec.

  • First pricing strategy at DAT. I introduced segment-level pricing research to set the 2025 budget and five-year plan. It informed broker list price increases of roughly 10–23% across the lineup (Select Broker ~$260 → $319), shaped packaging, and informed buy and partner calls.

  • Problems to solve, not features. Questions and readouts were framed around jobs to be done and the broker’s current solutions, grounded in our living experience map. Context is king; raw survey scores weren’t trusted alone. It’s important to note that customers often think about things in current state and not future state solutions, form factors, or the incremental path to get there.

  • 1,043 completes from a census of our broker network. I measured each surface three ways: top-2 box desirability, simplified Kano, and pairwise forced choice. Results reached about ±7–9% on needs at 95% confidence. Product scored effort per segment, and finance rolled segment tech spend into TAM, SAM, and SOM.

  • Built for real constraints. This was a first attempt on SurveyMonkey, with response rates under 2% and a small enterprise population. I chose methods that held up at that scale, labeled enterprise reads directional, and filled external gaps with the Bain engagement.

  • Trust carried the price. Carrier vetting and fraud scored as a must-have for mid-market and enterprise brokers, so the Carrier Management Suite became a lead part of the value story behind the new pricing. Its pieces were packaged into the higher tiers: CarrierWatch in Select Broker, and insurance certificates in Office Broker. The suite reached $4.93M MRR by Q4 2025, and the work built the case for better research tooling, including Pendo.

A survey like this captures how the market sees and ranks things today. Respondents can rank the needs they already understand, but they can’t price surfaces they haven’t seen or gauge a maturity that only shows up as you ship. So this artifact did one job: point the strategy at which markets to prioritize and roughly how hard. What to build inside each market lived downstream in PRDs and the product cycle. It was the foundation, not the floor plan.

It stayed a living input rather than a one-time deliverable. We shipped first iterations of the surfaces it prioritized and kept maturing them as the bigger ownership calls got made around them.

Method details

  • I tapped the entire broker population on our network rather than sampling, because at response rates under 2% a sample would have left most segments too thin. I pulled with SQL from our user database, segmented by company size and split between account holders and seat holders, staged the lists in Google Sheets, and fielded through SurveyMonkey.

  • I projected margin of error for each segment before fielding, at 95% confidence, using finite population correction for small segments and a clustering adjustment for multiple users from the same company. Across 1,043 completes, we landed at:

    • Needs and priorities (account holders and users pooled): about ±7% in 0–10 and 11–50, ±8% in 51–500, and ±9% in 501+.

    • Spend and pricing (one account holder per company): ±8% in 0–10, ±10% in 11–50, ±15% in 51–500, and ±18% in 501+.

    Margins describe precision, not representativeness.

  • The enterprise population is small, so margins stay wide no matter how hard you push email. I blended contacts that account managers personally invited with independent respondents from the email census, and removed the picked accounts from the SQL pull so no one got hit twice. Enterprise results went out labeled directional.

  • This was DAT’s first study of its kind, run on SurveyMonkey without advanced survey tooling. Van Westendorp and MaxDiff needed separate tools and budget we’d have had to advocate for, and price-sensitivity methods get statistically thin at B2B response rates. Top-2 box, simplified Kano, pairwise choice, and coarse price bands fit both constraints, and the coarser output let finance own pricing without a research bottleneck.

  • We had no way to survey brokers outside our network at the time. I folded in the $5 million Bain engagement to cover that gap, weighted as low-to-medium confidence because we couldn’t audit how the vendor collected its data.

    • Margins of ±15–18% on spend and pricing in the mid-market and enterprise segments

    • Friendly-contact bias in the account manager sample

    • The account holder isn’t always the buyer at larger brokers

    • A customer-only view of the market, with external signal from a source we couldn’t audit

    • Stated willingness to pay runs higher than actual

    • Margins of ±15–18% on spend and pricing in the mid-market and enterprise segments

    • Friendly-contact bias in the account manager sample

    • The account holder isn’t always the buyer at larger brokers

    • A customer-only view of the market, with external signal from a source we couldn’t audit

    • Stated willingness to pay runs higher than actual

    From the current section, I carried over three things:

    • the opening framing (which markets to aim at, built with the senior director of product)

    • the “foundation, not the floor plan” argument

    • the “living input” line

What this transfers to

The freight specifics matter less than the sequence, and the sequence travels to any fragmented platform landscape:

  1. Map the end-to-end experience.

  2. Find the high-trust decision points and ask who owns them.

  3. Partner until the seams hurt, then let penetration, product health, and margin pick your buy target.

  4. Own the surfaces you intend to train automation on.

  5. Decide form factor last, because what you can own is what you can build.

Where this is heading

The goal is to turn a load board into the operating system for American road freight, and it is happening in public. A traditional load board is a posting site: matches happen on the platform, but the deal, the payment, and the execution happen somewhere else. The next generation closes that loop and runs the whole transaction on one surface.

DAT has been assembling that surface node by node along the pipeline. It acquired TruckerTools for visibility and tracking, Outgo for payments and financing, and the Convoy platform for end-to-end capacity matching and execution, while phasing out its older TMS. Each one is a high-trust decision point the company chose to own rather than depend on, and each is integration-viable rather than a ground-up rebuild. That is the build/partner/buy framework running at company scale: own the surfaces that anchor trust and that you intend to automate on, integrate where you can, and consolidate the pipeline into a single operating layer.

The scale signals the stakes. As FreightWaves has reported, DAT's parent Roper carries roughly a $60 billion market cap and has put over $450 million into freight-tech acquisitions in a span of months, with the Convoy platform reported around $250 million. These are large bets on owning the matching layer outright, and they follow the same logic the experience map surfaced: the company that holds the end-to-end ownership position is the one that gets to define the form factor everyone else has to plug into.

Example in house build recommendation shipped : see the DAT Carrier Management Suite case study.

References and sourcing

The acquisition sequence (the nodes)‍ ‍DAT acquires Trucker Tools (closed December 17, 2024). DAT release: https://www.dat.com/company/news-events/news-releases/dat-acquires-trucker-tools-an-innovator-in-real-time-load-tracking-and-visibility and FreightWaves: https://www.freightwaves.com/news/dat-acquires-trucker-tools-leadership-speaks-on-tech-goals-for-2025

DAT acquires Outgo (announced May 15, 2025). Press release (Nasdaq/BusinessWire): https://www.nasdaq.com/press-release/dat-redefines-freight-payments-outgo-acquisition-2025-05-15 and FreightWaves, "DAT acquires Outgo, enters race to become dominant freight exchange platform": https://www.freightwaves.com/news/dat-acquires-outgo-enters-race-to-become-dominant-freight-exchange-platform

DAT acquires the Convoy platform (July 2025, reported around $250M). FreightWaves, "Load-matching wars escalate as DAT snaps up Convoy": https://www.freightwaves.com/news/load-matching-wars-escalate-as-dat-snaps-up-convoy

DAT corporate profile (the stack, in DAT's own words). https://www.dat.com/

"Roper Builds All-in-One Freight Platform Around DAT Ecosystem," Supplychain360 (July 2025).https://supplychain360.io/roper-builds-all-in-one-freight-platform-around-dat-ecosystem/

Pricing and packaging (the price story)
DAT broker plans, current list prices (Express $159, Select Broker $319, Office Broker $449; tier features including CarrierWatch and insurance certificates). DAT: https://go.dat.com/compare-broker-plans
Prior DAT One broker list prices (Express $145, Select Broker $260, Office Broker $395; undated listing). TrustRadius: https://www.trustradius.com/products/dat-one/pricing
Carrier Management Suite outcomes ($4.93M MRR, Q4 2025). See the DAT Carrier Management Suite case study.