Change orders don't start in the field. They start in the bid room — in a scope package that was too thin, a sub that was too vague, or an assumption that was never written down.
By the time the RFI lands, the damage is already done.
According to FMI's Construction Disconnected report, $31 billion in rework occurs every year in the U.S. — 26% driven by communication breakdowns and 22% by bad project data. That's not a field problem. That's a preconstruction problem.
This guide breaks down exactly where change orders come from, and what your pre-construction team can do before bid day to stop them.
Most change orders trace back to one of three root causes:
All three are created before the first shovel breaks ground. All three are preventable with tighter preconstruction habits.
Arcadis' 2025 Global Construction Disputes Report puts the average U.S. construction dispute at $60.1 million. And for six of the last nine years, the number-one dispute cause has been errors and omissions in contract documents. Not weather. Not owner changes. Errors and omissions — the kind that live in scope packages.
Change orders as a share of project cost average 8–14% on commercial work. On projects with weak scope definition, that number climbs above 25%. The gap between a well-scoped project and a poorly scoped one is measured in millions.
Scope gaps aren't random. They're produced by specific habits — some intentional shortcuts, some inherited from the previous job. The Scope Gap Playbook, built from interviews with 200+ GCs, identifies the patterns that repeat across firms and project types.
This is the most-cited scope anti-pattern, and the most dangerous. It transfers no actual understanding of the work. When a dispute arises, "as per plans and specs" means nothing — because both sides read those documents differently.
A scope package from a similar project feels like a head start. It's not. The sub lineup is different. The drawings are different. The exclusions need to be different. Copy-pasting buries last project's assumptions inside this project's contract.
One estimating manager at a Canadian ICI GC put it plainly: "If you miss anything, they'll bill it."
Scope sheets are scoping tools. When project managers use them to negotiate contract terms — payment schedules, lien waiver timing, insurance requirements — the actual scope gets squeezed. The result is a document that does neither job well.
A scope package reviewed at the last minute is a scope package that hasn't been reviewed. Gaps aren't caught. Conflicts aren't flagged. And anything missed on bid day becomes a change order on day 60.
These patterns show up across firm sizes and project types. Recognizing them is the first step to stopping them. The full breakdown — including trade-specific versions — is in The Scope Gap Playbook.
Industry averages are useful. Specific examples land harder. Here are real scope gaps — anonymized from operator interviews — that became change orders:
None of these started in the field. Every one of them started in a scope package.
The firms with the tightest margins do these six things consistently. They're drawn from the Eight Habits framework in The Scope Gap Playbook.
Scope packages built from boilerplate first miss what's actually on the drawings. A pre-construction lead at a Top-ENR Canadian GC described the standard: "It's descriptive — bread, put it on a plate, use the open jar… You have to get to that level of detail or else they'll just be like, 'you didn't tell us that.'"
Start with the drawings. Build the scope package from what's actually there. Use the boilerplate as a checklist for what you might have missed.
"Including all work described in the contract documents" is not a scope item. It's an invitation to dispute.
Scope packages that reduce change orders name the drawing number, the spec section, and the page. Every scope item should be traceable to a source document. If it can't be cited, it probably isn't scoped clearly enough.
The difference between bid-day buyout and pre-bid scope alignment is enormous. When subs know what's expected before they price, they include it. When they find out after, they bill it.
A director of pre-construction at a mid-market Southeast GC described the disconnect that creates change orders: "Pre-con is working in the scope sheet world and project management is working in the scopes of work." The fix is alignment — earlier in the process.
Exclusion lists tell subs what they're not doing. Clarification lists tell them what they are doing — and how.
Relying only on exclusions leaves gray area. If it's not explicitly excluded, an owner or owner's rep may argue it was included. Clarifications narrow that gray area and reduce the surface area for disputes.
A national mechanical sub doesn't need the same level of detail as a local specialty contractor on their first hospital project. Over-specifying wastes time. Under-specifying on a less-experienced sub creates gaps.
As the estimating manager at a Canadian ICI GC noted: "We have less subs who just kind of [operate on] a gentleman's agreement… they've become more quick to clarify that we're not including that one piece of scope."
Match the scope detail to the sub's track record. That calibration is what separates GCs with tight buyouts from GCs chasing change orders.
Every scope package should be reviewed before it goes out — not after the bids come in. The review should check for:
This is the checkpoint most firms skip. It's also the one that catches the $400K gaps before they become change orders.
Some trades generate far more change orders than others. Not because they're harder — because their scope boundaries are inherently ambiguous. The Scope Gap Playbook's trade-specific chapter breaks down the recurring gaps by division.
Generator field conditioning, motor starters, fire-rated louvres, and lighting controls are the most commonly disputed MEP scope items. The core issue: MEP trades have overlapping responsibilities, and the spec often doesn't resolve them clearly.
"The trench" is a classic example. Who owns the trench between mechanical and plumbing? If the scope packages don't say, both subs will assume the other. That assumption costs money.
Grouting base plates, slab assembly conflicts, and concrete pumping are the recurring gaps. Drawings often show a slab assembly that doesn't match the spec. The conflict gets discovered in the field — and becomes an RFI, then a change order.
Embeds and anchor bolts, hoisting and rigging, and miscellaneous metals are the gaps that hit hardest. Miscellaneous metals scatter across divisions — and no single sub feels responsible for the full scope.
Roof cover board, masonry-to-storefront flashings, caulking at interior vs. exterior joints, and fire-stopping are the recurring envelope gaps. These are especially high-risk on projects where the envelope scope crosses multiple sub trades.
Manual scope package development takes 30–40 hours per bid. That's not a guess — it's the baseline Provision measured across hundreds of GC projects. At that pace, most pre-construction teams can only do a thorough scope review on their best-fit pursuits. Everything else gets a lighter pass.
That's where scope gaps live.
Scope Agent generates a complete scope-of-work package from construction documents — drawings, specs, and addenda — in under 60 minutes. It reads the full project set, assigns scope to trades, flags conflicts between drawing sets, and identifies items that appear on drawings but aren't called out in the spec.
On a validated hospital project, Scope Agent captured 145 line items with 91.7% verified accuracy — with zero fabricated entries. That's not a surface scan. That's the kind of depth that catches the $300K lead-lined glass gap before bid day.
Provision has reviewed more than $100 billion in project value and processed more than 100,000 documents. The scope extraction accuracy on its internal validation dataset is 97%. Human estimators on the same exercise averaged 91.3% — at roughly four days per project versus under one hour.
This doesn't replace the estimator's judgment. It gives estimators a complete first draft to react to, instead of building from scratch. The time savings go back into sub-specific tailoring, pre-bid scope alignment, and the review checkpoints that prevent change orders.
The EllisDon case study shows what that looks like in practice — $1.8M in scope gaps identified before contract execution.
A scope package that reduces change orders has four properties:
A scope of work template is a starting point — not a substitute for this process. If you need a baseline, Provision's scope of work template is built for GC pre-construction workflows.
For GCs looking to tighten the full pre-con process — from bid to buyout — the Scope Gap Playbook is the most complete resource available. It covers the eight habits, the trade-specific gaps, and the contract language patterns that create change order exposure.
Most change orders trace back to scope gaps created at bid or buyout — not field surprises. The most common root causes are missing scope assignments, conflicts between drawings and specs, and assumptions that were never written down. Arcadis data shows errors and omissions in contract documents have been the top dispute cause for six of the last nine years.
Change orders average 8–14% of total project cost on commercial work, according to Navigant data republished by the AIA. On projects with weak scope definition at bid and buyout, that figure can exceed 25%. The difference between tight and loose scope packages is often measured in millions.
Before construction starts — ideally before subs are awarded. The preconstruction phase, specifically bid and buyout, is where scope is defined and disputes are seeded. Scope gaps caught before award cost nothing to fix. Gaps caught in the field cost time, money, and relationships.
A scope gap is a piece of work that appears in the project documents but isn't clearly assigned to any party in the subcontract. When that work needs to be done in the field, someone has to pay for it — and that's usually the GC, through a change order or a margin hit absorbed under "readily inferable" contract language.
MEP, concrete, steel, and envelope trades generate the most scope-gap-driven change orders. Recurring problem areas include generator field conditioning (MEP), slab assembly conflicts (concrete), embeds and rigging (steel), and roof cover board and flashings (envelope). These gaps appear repeatedly because scope boundaries in these trades are inherently ambiguous.
AI tools built for GC workflows — like Scope Agent — read the full project set (drawings, specs, addenda) and generate complete scope packages in under 60 minutes. That replaces 30–40 hours of manual work per bid and surfaces conflicts between document sets before subs are awarded. The time saved goes back into the review work that actually prevents change orders.
"Readily inferable" is contract language that requires a contractor to perform work that a reasonable professional would consider implied by the contract documents — even if it's not explicitly stated. GCs who absorb scope under this language lose margin without a change order. Subs are increasingly pushing back on this standard, making explicit scope assignment more important than ever.
Scope Agent generates a complete scope package from your project documents in under 60 minutes.
See Scope AgentMore Articles
