How many revision rounds should a redesign include, and what counts as one round?

Here is the test. If the studio can read the feedback once, act on all of it, and show you a new version, that was one round. If the studio has to stop halfway because a second email arrived contradicting the first, you have spent two.

So a round is defined by three things: one deliverable, one deadline, one voice. The deliverable is named, for example the homepage design or the built pricing template. The deadline is the date you agreed to send by. The voice is consolidated, meaning your team’s disagreements were settled before the message left your building, not after.

In our projects the count is rarely what runs out. What runs out is patience, on both sides, when a first round arrives as three separate replies that want three different hero sections.

• **A round:** twelve comments on the homepage design, sent Thursday, agreed internally, including two you have decided to drop.
• **Not a round:** four comments Monday, three more Wednesday, and your CEO's reply-all on Friday.
• **A round:** "We accept the layout. Change the headline, tighten the card spacing, replace the second image."
• **Not a round:** "Marketing likes option A, sales likes option B, let us know what you think."
• **A round:** comments on the deliverable currently under review.
• **Not a round:** comments on the navigation two weeks after the navigation was approved.

A fixed quote assumes a fixed amount of work. Review is the one part of a project a studio cannot bound on its own, because it depends entirely on how your side behaves. Capping rounds at two or three is the crude instrument that stands in for a rule nobody wants to write: please decide.

Said honestly, that cap also protects you. A studio with unlimited review has no reason to get the first version close, and every reason to send something safe and wait for you to design it by committee over eight weeks. A cap makes the first version matter to them.

Be wary of an unlimited revisions promise. It is priced in somewhere, usually in a higher day rate, a longer timeline, or a scope clause that is narrow enough to make most of your comments chargeable anyway. Unlimited rounds on a tightly drawn scope is worth less than three rounds on a generous one.

The late voice is the most common. Feedback goes out, work restarts, and then someone senior who was on holiday sees the design. Their comments may be right, but they arrive against a version that no longer exists, and the studio now reworks something it already reworked.

The reopened phase is the most expensive. Glasspage runs four milestones in order, Foundation, Content system, Page deployment, Launch ready, and each ends in a written approval. Feedback on typography during Page deployment is not a late comment on the current deliverable. It is a change to something that was signed off and built on, which is why it costs more than the same comment would have cost three weeks earlier.

The third is comments on the wrong deliverable. During a build review, half the notes are about copy. Nobody is wrong to want better copy, but the studio cannot act on it, the note sits in the thread, and the round ends without resolving. Copy notes belong to your side of the work, because you supply the final copy.

Do it in this order, because each step removes the reason the next problem happens.

One before and after, from a real pattern we see. The raw batch from a client team read: “Hero feels flat.” “Can we try a photo instead of the illustration?” “Love the illustration.” “Make the CTA bigger.” “Is the CTA even needed here?” That is five comments containing two direct contradictions and one question. Working on it would have produced a version somebody was guaranteed to reject.

Reconciled, the same batch became: “Keep the illustration. Increase the hero headline size one step. Keep the single CTA, make it the primary button style.” Three instructions, no contradictions, one version back. Same people, same opinions, one round instead of two.

1. Name the deliverable and the date. "Homepage and pricing designs, feedback by Thursday 5pm."
2. Collect from everyone who has a right to comment, and tell everyone who does not that they are informed, not consulted.
3. Reconcile conflicts yourself. Where two people disagree, one decision-maker picks, and the losing view is dropped rather than passed on as an option.
4. Sort what is left into change requests and questions, and mark anything that is a preference you can live without.
5. Send once, as one message. Anything that arrives after goes into the next round.
6. Confirm what you accepted, not only what you want changed. Silence on a section reads as unresolved, not as approval.

It is fine when the remaining notes are small and specific: a caption, a button label, a crop. Those are close-out items, and most studios absorb them because they cost minutes. It is also fine when you have simply changed your mind about something real, like swapping a testimonial section for a logo strip. That is a new request, and paying for it is the honest transaction.

It is a signal when round three is still arguing about the same thing round one argued about. That is not a client failing to decide. It usually means the brief never settled what the page was for, so every version can be criticised and none can be approved. The fix is not a fourth round. It is a thirty-minute conversation about the goal of the page, then one more version.

And when you are three rounds in and still need changes, buying an extra round is usually cheaper than fighting the scope clause. An extra round costs a known amount. A two-week argument about whether your comment was in scope costs your own time, the studio’s goodwill, and the launch date, and it often ends with you paying anyway.

Ask three questions of any proposal. How many rounds per deliverable, not per project, because two rounds across an entire site is not the same offer as two rounds on each design. What defines a round in writing. And what an extra round costs, stated as a number.

If the answer to the third question is “we’d have to look at it,” you have no way to price your own indecision, which is the thing rounds exist to price.

The count is the least interesting number in the contract. Two rounds collected properly will get you a site you are happy with. Five rounds arriving one email at a time will not.