This request shows up in two very different forms, and they need different responses. Sometimes a client asks for "unlimited revisions" explicitly, upfront, during scoping. Sometimes nobody ever said the word "unlimited," but a project starts behaving that way anyway, because nothing ever stopped it.
When it's asked for explicitly
Some clients request unlimited revisions genuinely believing it's a reasonable, even generous-sounding ask — "I just want to make sure we get it right" sounds collaborative, not extractive, from their side.
Reframe it around outcome, not round count. "I want to get it right too — in my experience, 2–3 focused rounds get us there better than an open-ended process, which tends to lose direction rather than sharpen it. Let's start there, and if we genuinely need more, we can talk about it." This isn't a refusal — it's steering toward what actually produces better work, which happens to also protect your time.
If they insist, price it accordingly. Unlimited revisions is a real, quantifiable risk to your time, and it's reasonable to price a project higher to account for it — either as a premium flat fee or as hourly billing beyond a certain point, explicitly stated.
There's research suggesting the bound actually serves the client too
Here's an angle worth bringing into this conversation that goes beyond "it costs me too much time": psychologist Barry Schwartz's well-known research on decision-making styles distinguishes between maximizers — people who, faced with a choice, compare every available option exhaustively before committing, and remain anxious even afterward about whether they picked correctly — and satisficers, who settle on something meeting their real criteria and move on without continuing to second-guess it. Across a substantial body of work, maximizers consistently report lower satisfaction, lower happiness, and lower confidence in their choices than satisficers do, despite typically investing far more time and effort, and sometimes landing on objectively comparable or even better outcomes. It's worth being honest that this finding isn't entirely uncontested — some later research questions how strong the link between maximizing and unhappiness really is — but the core pattern has held up across a wide range of replication and summary.
An unlimited-revisions arrangement structurally invites maximizer-style behavior: there's no defined stopping point, which means every version remains implicitly open to comparison against some imagined better one. Research on this pattern suggests that doesn't actually produce more confidence in the final result — if anything, it tends to produce less, because the process never generates the kind of closure that comes from deliberately choosing to stop. A bounded revision process, counterintuitively, may do the client a genuine favor beyond just protecting your schedule: it gives them a real stopping point, which the research suggests correlates with walking away more satisfied with what they got, not less.
There's a related finding worth knowing, since it explains why this request often surfaces specifically on a client's highest-stakes project rather than evenly across all their decisions: people aren't maximizers or satisficers uniformly across every choice they make — the tendency is domain-specific, and people often shift into maximizer mode specifically for decisions that feel higher-stakes or harder to reverse. A client who's perfectly decisive about smaller choices may become genuinely unable to settle on a direction for something they perceive as more consequential — a brand identity, a flagship campaign — which is exactly the kind of project where "can we just keep revising until it's perfect" tends to show up.
Recognizing the pattern while it's happening
A few signs tend to show up when a revision process has drifted into the kind of open-ended comparison the research above describes, worth watching for before round 5 or 6 arrives:
- Feedback starts referencing earlier versions again, not just the current one. "Actually, can we go back to something like round 2" after round 4 has already moved past it is a strong signal the client isn't converging toward a decision — they're still comparing.
- Approval language gets more tentative over time, not less. "This works, I think" or "let's go with this for now" after several rounds often indicates declining confidence, not growing satisfaction — the opposite of what more rounds are supposed to produce.
- New criteria keep appearing that weren't part of the original brief. Each new round introducing a genuinely new consideration, rather than refining against the same original goals, suggests the comparison has become open-ended rather than converging.
None of these are a sign the client is being difficult — per the research above, it's a well-documented pattern that shows up independent of anyone's intentions. It's worth naming directly when noticed: "I want to make sure we're converging, not just generating more options to compare — should we lock in a direction from what we have so far?"
When it happens by drift, not by request
More often, "unlimited" isn't a term anyone used — it's just what happens when a revision count was never specified and nothing in the process enforces one. Round 4 arrives, and there was never a clear moment to say "we're past what's included," because there was never a defined "included" in the first place.
Introduce a count retroactively, gently. "Just so we're aligned going forward — we've done 3 rounds so far. Happy to keep refining, and future rounds beyond this point would be [$X each / billed hourly]." This isn't punishing the client for something never stated; it's establishing structure going forward.
For the next project with the same client, define it upfront. Once a revision limit exists in writing and is visible throughout the project, the drift that created "unlimited" in the first place stops happening by default — and, per the pattern above, a client who previously spiraled through an open-ended process on one project is often visibly relieved to have a defined structure the second time around, once they've felt what the alternative cost them.
A worked example
A brand designer agrees to "unlimited revisions" on a logo project for a first-time client launching their business — a genuinely high-stakes decision from the client's perspective, the kind of choice the domain-specific research above suggests is most likely to trigger maximizer-mode behavior. Round 3 produces a version the designer considers strong and on-brief. The client approves it, tentatively, then two days later asks for "just one more look at a different direction, to compare" — not because anything is wrong with round 3, but because nothing ever closed the door on comparing further.
By round 7, the client is revisiting elements from round 2 that round 3 had already moved past, visibly less confident than they were after round 3, not more. The designer, exhausted, finally proposes locking in round 3 as final. The client agrees with visible relief — not because round 3 was objectively better than rounds 4 through 7, but because a real stopping point had finally been introduced from outside, which the open-ended process itself never provided.
A bounded process — 2 rounds, clearly stated upfront — would very likely have landed on something close to round 2 or 3 regardless, arrived at faster, with the client never given the opening to spiral into comparing an ever-expanding set of alternatives against each other in the first place.
Why this keeps recurring without a system
The core issue in both cases is the same: revision count is either not defined, or defined but not tracked anywhere both sides can see. A SwiftSignoff milestone has a revision limit built in from the start, visible to the client throughout — "1 of 2 revisions used" — so neither the explicit "can we make it unlimited" ask nor the quiet drift into round 4 can happen without becoming visible first. And per the research above, that visible bound isn't purely a protective measure for the freelancer — it's a structural nudge toward the kind of defined stopping point that tends to produce more client confidence in the outcome, not less.
FAQ
Is "unlimited revisions" ever a reasonable thing to offer?
For a very small, low-stakes project, occasionally. For most work, it's a real liability — the time cost is open-ended while the payment isn't, and research on decision-making suggests it may not even serve the client's satisfaction as well as a bounded process would.
How do I price a project if a client insists on unlimited revisions?
A meaningful premium over your normal rate (many freelancers use 25-50% or more) reflects the real, open-ended risk being taken on.
What if a client says other freelancers offer unlimited revisions?
It's fine to hold your position: "That's a model some people use — for the quality and turnaround I aim for, a defined process with a couple of focused rounds tends to work better for both of us."
Does a revision limit make me seem less collaborative?
Generally not, if framed around getting to a better result faster rather than around limiting the client — most clients respond well to structure once it's explained as being in service of the work, not a restriction on it.
Could unlimited revisions actually make a client less happy with the final result, not more?
Research on decision-making styles suggests this is plausible, though not fully settled. People who keep comparing an unbounded set of options before committing ("maximizers") tend to report lower satisfaction with whatever they eventually choose than people who stop once their criteria are met ("satisficers"), even when the choices themselves are comparable in quality. An open-ended revision process structurally invites the first pattern.
"Unlimited" is a real cost with no price attached — and it may not even make the client happier with what they get. Put a visible number on it instead. See how revision limits work.