The Clause Nobody Writes Until It Is Already a Problem
Almost every agency that offers “unlimited revisions” as a selling point learns the same lesson eventually, usually from one specific client. It sounds generous in a pitch, a confidence signal that the agency will keep working until the client is happy. In practice, without a defined boundary, it behaves less like a service guarantee and more like an open invitation, and the account team is the one who discovers where that invitation ends.
The clause that fixes this almost never gets written proactively. It gets written the week after a client asks for a ninth round of changes on what was supposed to be a two-round deliverable, at which point the agency finally sits down to define, in writing, what it should have defined before the engagement started.
Why “Unlimited” Behaves Exactly Like Scope Creep
Scope creep is the continuous, generally harmful growth of a project’s scope after it begins, and one of its most common causes is what gets called the low cost of change trap: each individual small request looks reasonable and cheap to accommodate on its own, so nobody on the delivery team pushes back on any single one. The requests accumulate into a real cost only when someone finally adds them all up, usually well after the damage is done.
An unlimited-revisions offer removes the one mechanism that would normally interrupt that pattern, a defined limit that forces a conversation before request number six. Without it, “just one more small change” can repeat indefinitely, each instance individually reasonable, the sum of them nowhere close to reasonable.
The Missing Piece: What “Acceptable” Means
A standard statement of work is understood to cover ten components, and one of them is acceptance criteria, a defined standard for what counts as a completed, approved deliverable. Most agencies that get burned by unlimited revisions never had that component spelled out in the first place. Without a defined “what does an acceptable first draft look like” standard, a revisions clause has nothing to measure against, no line a ninth request can be shown to have crossed.
That absence is the actual root cause, more than the word “unlimited” itself. A revisions clause with a real cap but no acceptance standard still invites the same argument over whether the current draft counts as done. The acceptance criteria is the piece that makes any revisions clause enforceable at all.
The Conversation That Happens After the First Bad Experience
For most agencies, the fix arrives as a retroactive conversation with an existing client rather than a clean policy rolled out in advance: a defined number of revision rounds per deliverable, and a written acceptance standard for what a first draft has to hit before revisions even start counting. That conversation is uncomfortable precisely because it happens mid-relationship, after a pattern has already been set and a client has already gotten used to open-ended access to changes.
It is still worth having. A late fix that holds is better than no fix at all, and most clients, confronted with a clear, reasonable standard rather than a vague complaint, adjust their expectations without the relationship ending over it.
Writing the Clause Going Forward
For every engagement after that first hard lesson, the clause itself is straightforward to write once the underlying problem is understood: a specific number of revision rounds included in the scope, a defined acceptance standard for what a draft has to satisfy before those rounds even begin counting, and an explicit statement of what happens once the included rounds are used, an additional round at an hourly rate, or a new change order entirely.
The goal is not to punish clients who ask for changes, reasonable revision requests are a normal part of collaborative creative work. The goal is to make sure the agency and the client are working from the same definition of where “reasonable” ends and a new scope of work begins.
Balancing the Clause Against the Relationship
A hard cap enforced without any flexibility can feel adversarial exactly when a client relationship needs goodwill the most, so most agencies that write this clause well pair it with a practical release valve: a documented process for approving one extra round outside the cap when the situation genuinely warrants it, logged as an exception rather than treated as evidence the cap does not really apply.
That combination, a real limit plus a documented, occasional way around it, tends to hold up better over time than either an unlimited promise or a rigid rule with no room for judgment.
Sources
The external data in this article draws on the sources below. Figures described in the text as estimates or industry triangulations are directional and are not attributed to a single dataset.