Why so many change orders hit a single project is rarely a mystery once you actually trace them back. Owners tend to talk about change orders as if they're weather — unpredictable, something to budget around rather than something to prevent. That framing holds for two of the three categories that drive them. It does not hold for the third, and that third category is usually the biggest preventable slice on the set: design errors and coordination gaps that were sitting in the documents, readable, before a single crew showed up.
That distinction is worth taking seriously, because it changes what an owner should actually do about change order volume. Two of the three root causes genuinely can't be eliminated by reading the drawings more closely. The third one can — and understanding why it keeps happening is the first step to catching it before it turns into a line item.
The Three Places Change Orders Actually Come From
A January 2025 report from the U.S. DOT's Project Delivery Center of Excellence, produced by the Volpe National Transportation Systems Center, grouped the root causes of construction change orders into three categories: the quality of technical project development work, the influence of organizational culture on how issues get raised and resolved, and uncertainty introduced by financial forces outside the project team's control. That framing lines up with what shows up on any large capital project, public or private.
Only the first category — the quality of the technical work itself — is something a document-level review can act on before the set ever gets priced. Organizational culture and financial uncertainty are real, but they aren't fixable by reading a drawing set more carefully. The technical-quality category is different: it's made up of things that are already wrong on paper, sitting in the issued set, waiting for someone to notice.
A Texas Department of Transportation study tracking change orders across eleven years of highway projects found that 31% of the total change orders — averaging $45 million a year — were tied specifically to design errors and omissions. That's not a rounding error in the total change order picture; it's close to a third of it, on a category of projects with far more design review scrutiny than most private commercial work ever gets.
Why the Design-Error Slice Keeps Slipping Through
If design errors and coordination gaps are genuinely catchable on paper, the obvious question is why so many change orders in that category still happen. The answer isn't that nobody reviews the set — it's how the set typically gets reviewed.
Most quality control on a construction document set happens discipline by discipline. The structural engineer checks the structural sheets against structural standards. The mechanical engineer checks the mechanical sheets against mechanical standards. Each pass can be thorough and still miss the conflict that only exists between two disciplines' sheets — a duct routed through a beam, a note on one sheet that contradicts a detail on another, a spec section calling for an assembly the drawing doesn't match. None of those show up in a single-discipline QA/QC pass, because each sheet, read on its own, is internally correct. The problem only exists in the gap between sheets — which is exactly the ground covered in why "it passed QA/QC" doesn't mean the set is coordinated.
Add schedule pressure on top of that gap and the pattern gets worse, not better. Cross-discipline coordination checks are the first thing compressed when a set has to move to bid on a fixed date — not because anyone decided coordination didn't matter, but because nobody explicitly owns catching the conflicts that live between disciplines instead of inside one. That ownership gap, more than any single mistake, is why so many change orders in the design-error category keep making it all the way to the field before anyone catches them.
Where the Preventable Conflicts Actually Sit on the Set
The specific conflicts that drive this category aren't random. They cluster in a small number of predictable spots:
- MEP-to-structural clashes — a duct, conduit run, or pipe routed at an elevation that conflicts with a beam or column drawn on a different sheet, as walked through in detail in how a single missed MEP clash turns into a $400,000 change order.
- Spec-vs-drawing contradictions — a written spec section calling for one assembly while the paired drawing shows a different one.
- General notes that contradict a detail — both defensible read alone, impossible to build together.
- Schedule mismatches — a door schedule that doesn't match its hardware spec, a panel schedule that doesn't match its one-line diagram.
Every one of these already exists on the issued set before a shovel goes in the ground. None of them requires a field surprise to happen — they're sitting on paper, waiting to be read by someone whose job is specifically to check one discipline's sheets against every other discipline's, not just against itself.
Catching It Before the Set Gets Priced
Once the design-error slice is understood as a coordination gap rather than an unlucky field discovery, the fix stops being mysterious: someone has to read the full set — drawings and specs, every discipline against every other discipline — before it goes to bid, specifically looking for the conflicts a single-discipline QA/QC pass is structurally unable to catch. That's a different exercise than a code check or a peer review of one discipline's own work; it's a cross-discipline read with the explicit job of finding where the sheets disagree with each other.
Timing changes what that catch is worth, too. A conflict found before the set is priced is still just a redline — cheap to fix, nobody's contract is built around the wrong version yet. The same conflict found after subcontractors have already bid the set as drawn means reconciling what was priced against what actually has to get built, which is a slower and more expensive version of the identical fix, as covered in pre-bid vs. IFC review timing. And once the exposure sitting in a given set is treated as a real number instead of a fixed cost of doing business, it becomes something an owner can act on directly — the actual math for that is laid out in change order math: calculating real exposure before you bid.
Preempt Global's document review is built around exactly that gap: reviewers who've drawn these sets themselves read the full issued set — every discipline against every other — within 48 hours, backed by a $50,000-exposure-or-free guarantee. That guarantee only makes sense as a business model because the design-error slice described above isn't theoretical. It's reliably sitting in sets of this size and complexity, in the same handful of places, waiting to be read before it becomes a change order instead of a redline.
Key takeaways
- Change orders trace back to three root categories: technical document quality, organizational culture, and financial uncertainty — only the first is fixable by reading the set more closely.
- A TxDOT study spanning eleven years found 31% of all change orders, averaging $45 million a year, tied specifically to design errors and omissions.
- Most QA/QC happens discipline by discipline, which structurally misses conflicts that only exist between two disciplines' sheets — the exact gap a cross-discipline review is built to close.
- The preventable slice clusters in predictable spots: MEP-vs-structural clashes, spec-vs-drawing contradictions, conflicting general notes, and schedule mismatches.
- The same conflict is cheapest to fix as a pre-bid redline and gets progressively more expensive the later it's caught — timing, not luck, drives most of the cost difference.
Frequently Asked Questions
Why so many change orders happen on large construction projects?
Most change orders trace back to one of three root causes: owner-requested scope changes, unforeseen field conditions, and design errors or coordination gaps already sitting in the issued set. The first two are genuinely hard to predict. The third is a document-level problem, and it's consistently the largest single preventable category — a TxDOT study found 31% of all change orders over an eleven-year span were tied specifically to design errors and omissions.
Can better QA/QC alone stop most change orders?
Not on its own. Most QA/QC checks a discipline's sheets against that discipline's own standards, which structurally can't catch a conflict that only exists between two disciplines — a mechanical duct routed through a structural beam, for example. Passing QA/QC confirms each sheet is internally correct; it doesn't confirm the sheets agree with each other.
What's the difference between a change order caused by design errors and one caused by unforeseen conditions?
A design-error change order comes from a conflict already sitting on paper in the issued set — something readable and catchable before the set goes to bid. An unforeseen-condition change order comes from something the drawings couldn't have shown, like a buried utility or subsurface soil condition discovered during excavation. Only the first category is something a pre-bid document review is scoped to prevent.
When is the best point to catch a design-error conflict before it becomes a change order?
Before the set goes out to bid, while the conflict is still just a redline nobody has priced around. The same conflict caught after subcontractors have already bid the set as drawn requires reconciling what was priced against what actually has to be built — a more expensive, more contentious version of the identical fix.