When the initial concept is reopened in the third revision, the issue is often not the number of rounds. The team has not agreed which decisions belong at which stage. I organise revisions around purpose before counting them.
First round: are we solving the right problem?
Review the direction, structure and central narrative. Correcting punctuation is premature if an article’s main argument is wrong. Fine visual details can distract from a page whose information order does not work.
Alternatives can be discussed here, but the selected direction and its rationale should be recorded by the end. Unresolved uncertainty should not quietly move into the next stage.
Second round: is the direction well executed?
Now the focus moves to flow, language, visual relationships and detail. A new idea may be appealing, but changing the agreed direction should be recognised as a separate decision. A small revision should not conceal a new assignment.
Feedback belongs in one consolidated place, with contradictions resolved before delivery. Actionable creative feedback makes each round useful rather than merely reducing their number.
Final round: is it ready to publish?
Check names, dates, links, language and output accuracy. This is the stage for preparing the accepted work correctly, not introducing an unrelated stylistic experiment. Genuine errors still need correction; they should be distinguished from preferences.
An unsupported claim discovered late cannot be ignored because the direction was approved. Stages help organise judgement; they do not protect mistakes from scrutiny.
Keep a visible outcome
A concise decision summary, current version and unresolved items are enough. It should be possible to see how the revision responded to the previous comments.
I want revision to reduce uncertainty step by step, rather than become an indefinite search for personal approval.