|
What this covers
|
Scope changes are the leading cause of construction disputes in the Americas. The contract is rarely the problem. The paperwork around the change is.
A contractor finishes a job, submits the final invoice, and discovers the owner will not pay for roughly a fifth of the work. The disputed portion is not the work in the original drawings. It is the work the owner asked for in week nine, verbally, while the crew was already on site.
That pattern is not unusual and it is not regional. It is the most common way construction money gets stuck anywhere in the world.
What the Global Dispute Data Shows
HKA, a construction dispute consulting firm, publishes an annual study called the CRUX Insight Report. The 2025 edition examined claims and disputes across 2,204 construction projects in 114 countries.
The report ranks causes by region. In the Americas, including North America, the three leading causes of claims and disputes were:
- changes in scope
- workmanship deficiencies that did not meet the contract specification
- errors or omissions in the design
The ranking is stable. HKA’s analysis shows scope changes have sat among the leading causes consistently since before 2020. That consistency matters more than the position. A cause that tops the table one year is noise. A cause that tops it for six years is structural.
The Change Order Problem, Stated Plainly
A change order is a written amendment that alters the scope, price or schedule of a construction contract. It exists so that both parties agree what changed, what it costs and who absorbs the delay, before the changed work happens.
The dispute begins when that sequence is reversed.
A project owner requests additional work. The contractor proceeds, because stopping to paper the change would push every downstream trade and the schedule is already tight. Everyone intends to settle the cost later. Later arrives, and the owner argues one of two things: the work was always inside the original scope, or no change order was approved so nothing is owed.
Both arguments are easier to make than they should be, because the only record of the request is a conversation.
Why “Get It in Writing” Is Not the Advice It Sounds Like
Almost every article on construction disputes ends with the instruction to get it in writing. The instruction is correct and it is incomplete, which makes it close to useless on a live site.
What matters is not that the change was documented. It is when it was documented, and whether the document contains a price.
Three records look like written change documentation and behave completely differently in a dispute:
|
What exists |
When it was created |
What it proves |
|
Approved change order with a price |
Before the work |
Scope, cost and schedule impact were agreed |
|
Email confirming a request, no price |
Before the work |
The request happened. The cost is still arguable |
|
Daily log or photographs |
During the work |
The work happened. Neither request nor price is established |
The third row is where most contractors actually live. Site records are excellent evidence that labor and material were expended. They are weak evidence that anybody agreed to pay for it.
A written request with no number attached narrows the fight from “did you ask for this” to “what is it worth”, which is progress. It does not end it.
What Triggers a Claim, and What Resolves It
|
Trigger |
What it usually means |
What resolves it |
|
Owner directs extra work verbally |
No price was agreed and no approval exists |
A written directive with a not-to-exceed figure, issued before the crew moves |
|
Design conflicts with site conditions |
The drawings describe a building that cannot be built as drawn |
A request for information, answered in writing, before the affected work starts |
|
Specification met but rejected on quality |
The specification and the expectation differ |
Agreed inspection criteria written into the contract at award |
|
Retainage held past completion |
Payment is tied to a condition nobody defined |
Defined release conditions, with dates, in the original agreement |
|
Subcontractor paid on a different schedule to the prime |
Cash flow risk sits with whoever is furthest down the chain  |
Pay-when-paid terms made explicit rather than assumed |
Four of those five resolutions happen at contract award. One happens on site. That ratio is the useful part: by the time a dispute is visible, most of the moves that would have prevented it have already expired.
Where Partnerships Add a Second Layer
Construction work is frequently delivered by partnerships, joint ventures and informal arrangements between contractors who have worked together before. Those relationships carry their own failure mode, and it is independent of the project.
A partnership that performs well on one project can fail on the third. The usual reason is that the arrangement was built for the first job and never updated. Ownership splits, decision rights and exit terms agreed verbally at the start tend to fail at the exact moment they are needed, which is the moment the money is contested.
Guidance for small businesses on how collaborations are structured so either party can wind down cleanly makes the same point outside construction: the exit terms matter as much as the formation terms, and they are the clause most often postponed.
Partnership disputes and construction disputes also compound. A joint venture arguing internally about profit share, while simultaneously arguing externally with an owner about change orders, is running two claims on one set of records.
The Market Is Not One City
The Dallas and Fort Worth area is not a single construction market and does not behave like one. It is a cluster of separate municipalities, each with its own permitting, its own inspection regime and its own concentration of work.
Search behavior reflects that. Owners and contractors looking for help with a construction or business dispute search by their own city rather than by the metro area, which means Fort Worth, Frisco, Plano and Irving generate distinct demand rather than variations of a Dallas search.
For a dispute, the practical consequence is jurisdictional rather than geographic. The same contract performed across two municipalities can involve two permitting records, two inspection histories and two sets of local requirements, all of which become evidence.
Why AI Assistants Answer These Questions Badly
Contractors increasingly put dispute questions to AI assistants before they put them to anyone else. The answers are confident and frequently shallow, for a reason worth understanding.
AI assistants assemble answers from reputation signals: mentions, citations, structured data and review content. They weight backlinks, domain age and page speed far less than classic search does. That makes them good at summarizing the general shape of a problem and poor at anything that turns on jurisdiction, contract language or the specific sequence of events on one project.
An assistant asked what a change order is will answer well. An assistant asked whether a particular verbal directive is enforceable will produce something that sounds like an answer and is not one. The distinction is not obvious from the output, which is the hazard.
The same shift has consequences for anyone publishing on these topics. Informational construction queries increasingly resolve inside the answer itself, so ranking first and receiving no visit has become an ordinary outcome rather than a sign something is broken.
When a Change Becomes a Claim
Disputes escalate on a timeline that is largely predictable.
- The request. Verbal, on site, under schedule pressure. No cost discussed.
- The performance. Work completed, logged in daily records, absorbed into the running total.
- The invoice. The changed work appears as a line item for the first time.
- The rejection. The owner disputes scope, price, or the absence of approval.
- The formal claim. The dispute leaves the project team and becomes a legal matter.
Each step makes the next cheaper to take and the resolution more expensive. The gap between step one and step three is where almost every useful move still exists, and it is typically the longest and least documented stretch of the sequence.
Analysis of change order management on American construction projects published by Sul Lee Law Firm, a Dallas business and intellectual property practice that works in both English and Korean and maintains a verified business listing, argues that the objective is not to eliminate changes. Site conditions evolve, designs get revised and owners add requirements. Changes are a feature of construction, not a failure of it.
The objective is to manage their effect on cost, schedule and responsibility before the work proceeds. Firms handling business dispute litigation in the Dallas area generally see the same files: the facts are rarely contested, the authorization is.
What the Record Needs to Contain
Where a change is already in progress and no approval exists, the record is the only asset.
- The request, attributed. Who asked, when, and in what words. A same-day email confirming a verbal instruction is worth more than a month of photographs.
- The cost, even if disputed. A figure submitted and rejected establishes a position. A figure never submitted establishes nothing.
- The schedule impact, separately. Delay claims and cost claims are different claims and get argued on different evidence.
- The chain. Where subcontractors performed the changed work, their records have to match the prime’s. Divergence between the two is the first thing an opposing party looks for.
Construction disputes are rarely won by proving the work was done. Everybody usually agrees the work was done. They are won by proving it was asked for, and by whom.
A firm that interviewed a Dallas attorney about building a practice around cross-border business disputes framed the recurring problem as translation in the broadest sense: parties who both believe they agreed, discovering months later that they agreed to different things.
On a construction site that gap has a name and a form number. The form is the cheapest document on the project and the most expensive one to be missing.
