If you're erecting steel today, you're probably already working on data centers, fabs, or large industrial capital projects. Those markets have changed the client mix significantly over the last few years. And the owners in those markets think about structural steel scope buyout differently than the institutional or commercial owners you got used to working for.
I don't mean that in a vague "they care about innovation" way. I mean it practically. These sophisticated owners are brining their company procurement practices into construction, whether we like it or not. They are moving toward evidence-based selection. And that shift is already showing up in how erectors are being evaluated even if it doesn't always look that way on the surface.
Here's something most people in the industry don't say out loud: on the vast majority of owner teams, almost nobody actually understands construction, let alone structural steel.
There's usually one person (maybe few) who really gets it. Who understands sequence, who understands what it takes to work through coordination pressure on a complex project, who can actually evaluate your plan rather than just your presentation. That person becomes your anchor. You build trust with them. You walk them through your plan. You make sure they understand how you operate, because they're the one translating your capability to the rest of the room, most importantly, when you are not there.
And once the decision moves beyond them to legal or procurement it stops being about how the work will actually get done. It becomes about how your proposal reads. How your team presented. How recognizable your name is.
A lot of winning work in steel today comes down to keeping that one person on your side.
Owners know this is a gap. And they've been trying to patch it.
Look at how proposals are structured today: you're seeing more technical scoring matrices than you used to. Approach. Methodology. Team structure. Execution plan. Each scored on a rubric. It looks more objective. In practice, it's still mostly a function of how clearly you present and how effectively someone advocates for you internally. If you're good at proposals, you can still shape the outcome.
But those scoring systems are a transition mechanism, not a destination. Owners are using them because they want something more reliable than pure narrative, and they don't yet have a better underlying signal.
That's what's changing.
Over the next several years, the visible scoring frameworks will start to fade. Not because owners stop caring about technical capability, but because the evaluation moves earlier in the process and further behind the scenes.
Instead of a committee scoring your proposal, you'll have internal (likely AI-driven) risk assessments running before you ever get to an interview. Systems that pull from structured performance data how you've actually executed on comparable work, how consistent that performance was, how your outcomes compare when conditions were similar.
Those systems won't be reading your narrative. They'll be analyzing your record.
When that happens, the ability to influence the outcome through presentation shrinks significantly. There's no room to frame the story, highlight selectively, or rely on your anchor to interpret your strengths. The system can see the data. Or it can't, if you don't have any.
For a long time, winning work in steel has depended on relationships, reputation, and the ability to tell a compelling story about your capability. None of that goes away overnight.
But the industry is in the middle of a real shift. The owners writing the biggest checks are getting better tools for evaluating performance, not just presentation. And when those tools mature, the erectors who've been building a credible performance record will be in a fundamentally different position than those who haven't.
That's not a distant scenario. It's already underway.
The obvious objections
I've made this argument to enough people in the industry to know what pushes back. Here's where I think the skeptics are wrong and where I think they have a point.
Construction procurement never really changes.
It's a fair observation historically. But look at what happened to logistics.
Amazon didn't just change how they bought delivery services. They built their own logistics network from the ground up, optimized entirely around performance data. Delivery times, route efficiency, cost per package, contractor reliability. They measured everything. And the incumbents who couldn't match that standard found themselves squeezed out. UPS just announced it's cutting Amazon volume by more than 50% by mid-2026 and eliminating roughly 20,000 positions as a result. FedEx, which walked away from Amazon in 2019 thinking it didn't need the volume, is now in the middle of a multi-year transformation to reinvent itself as a data-driven logistics platform. Amazon went from 1.7 billion packages delivered in 2019 to 6.1 billion in 2024. They are now the largest domestic parcel carrier in the United States.
That didn't happen because Amazon changed a procurement policy. It happened because they had better data about who could perform, at what cost, under what conditions and they built a system around it.
The owners building data centers and fabs today are the same kinds of organizations. They've already done this in logistics, in equipment procurement, in cloud infrastructure. Construction is still the category where performance is evaluated through narrative. That won't hold.
Raw production data doesn't capture context. A slow crew might have been working behind a delayed concrete pour.
This one is legitimate, and it's worth being precise about. Decontextualized data is a weak signal. An erection rate without site conditions, coordination environment, or design completeness attached to it doesn't tell you much. But that's an argument for better data capture, not against data-driven evaluation. The systems that will matter aren't the ones that record outputs in isolation — they're the ones that capture conditions alongside performance. That's actually a harder problem than most people building in this space appreciate.
The best erectors are the busiest. They have no incentive to instrument their work.
In the short term, maybe. But consider what happened to the GCs who made a deliberate bet on technically complex owners twenty years ago.
Hoffman Construction and Austin Commercial were not the largest contractors in their markets when they started pursuing semiconductor fabs, data centers, and advanced manufacturing. They weren't the obvious choice. The bigger, more established names had the relationships, the reputation, and the bonding muscle. What Hoffman and Austin had was a track record of execution on technically demanding work and owners who could actually see it.
Over twenty years, that positioning compounded. They didn't just grow. They became the preferred builders for the most demanding clients in the country, in the fastest-growing construction segments in the world. Not because they were biggest. Because they were verifiably capable.
That's the path that's opening up for erectors right now. The biggest names in structural steel got big on relationships and reputation built over decades. That's not easy to replicate. But being the erector who shows up with a credible production record when the owner is trying to evaluate capability directly that's a different kind of advantage. And it's available to anyone willing to build it.
You're describing a future that benefits Versatile. Of course you believe it.
Fair. We build the infrastructure that generates this data, so I have an obvious interest in owners valuing it. I'd rather acknowledge that directly than pretend the conflict doesn't exist. What I'd ask is that you separate the argument from the source and evaluate it on its own terms. The shift toward evidence-based procurement is happening independently of Versatile we didn't invent the trend, we're building for it.
The risk lives in the 20%. The unusual connections, the tight site, the owner who changes scope. Historical data doesn't predict that.
This is probably the strongest objection, and I don't want to dismiss it. Data won't replace judgment on the hard 20%. It was never going to. What it does is change how you get to that conversation. Right now owners spend most of their evaluation energy trying to assess the 80%. The baseline capability question that production data can actually answer. If that question gets resolved faster and more reliably, the human judgment gets concentrated where it actually matters. That's a better outcome, not a worse one.
Contractors adopted BIM and visual planning because it helped them. Data transparency helps the owner. Different incentive, different curve.
The incentive structure is different, that's true. But the framing of "this helps the owner" misses something. The erectors who build a credible production record aren't just giving owners a better tool for evaluation, they're giving themselves a better tool for winning work. Right now the erector with the strongest actual performance often loses to the erector with the strongest proposal. Data changes that dynamic in their favor. The incentive exists. It just hasn't been made visible yet.
I'm not going to let owners watch everything I do on a job.
You may not have a choice. More importantly, you may not realize it's already happening.
On virtually every significant data center project today, owners and GCs require steel erectors to document daily production as a condition of the contract. Piece marks installed. Location and sequence. Yard inventory reconciled against the fabricator's bill of materials. Progress tied to the lookahead schedule. Schedule of values draws, earned value reporting, and payment applications all depend on piece-mark-level data. Miss your reporting requirements and you risk delayed payment, schedule disputes, and liability exposure when the project goes sideways.
You're already reporting. Every day. On every major project.
Here's what most erectors haven't thought through: that data isn't just going into a binder. It's going into the owner's systems. And those owners, the hyperscalers, the industrial developers, the data center REITs are building exactly what Amazon built for logistics. A performance database. A structured record of who showed up, how fast they moved, and how their production held up under pressure.
Amazon didn't tell FedEx and UPS why they needed the reporting. They just kept measuring.
The same dynamic is coming to structural steel (and other trades). At some point, an owner is going to bring up your production history the one you've been building for them, on their projects, through your contractual reporting obligations.
If you think this stops at structural steel, it doesn’t. Meta (Facebook) is now applying the same approach to concrete. They’ve developed AI models that design and optimize concrete mixes based on real performance data toward continuous, data-driven optimization.
Their AI-designed mix reached full structural strength 43% faster while also reducing cracking risk. They’re open-sourcing the data and embedding these models directly into production workflows, where they continuously improve as more field data comes in.
Think about what that means.
The same owners:
are the ones buying your steel scope.
This is not theoretical. This is how they already operate.
You're reporting your production capability. The process has started.