Turn a Tracked-Changes Treatment Draft Into a Deliberately Clean Handoff

You set the view to No Markup, the struck-through words disappear, the page looks like a finished treatment, and you save it as the clean version. Nothing has been decided. Every revision is still sitting in the file, waiting for whoever opens it next to see something different than you did.

That is the trap this article is about. The word "clean" in a clean handoff does not mean a quiet-looking page. It means a saved copy in which every tracked change has been accepted or rejected on someone's authority, every comment has been handled on purpose, and you have reopened the exact file you intend to send to confirm both. Hiding markup changes what you see. It does not change what the document contains.

The mechanics come from Microsoft's support documentation on accepting and rejecting tracked changes in Word: hiding markup does not remove tracked changes, changes need an explicit acceptance or rejection, comments are deleted through a separate operation, and some controls act only on the reviewers currently shown. No Word document was opened or edited in writing this piece, and the controls differ between desktop and mobile Word. Confirm the exact labels in your own version before you commit anything.

Keep the version you were given, and work somewhere else

Before you change a single revision, decide three things and write them down.

Which file is authoritative — the one whose wording the approval actually refers to. Usually that is the version that was circulated, not whichever copy happens to be closest to hand.

What the recipient expects to receive. A treatment that a designer will set is a different deliverable from a PDF someone will read on a phone, and the file you send should be the one that matches the ask.

Who can decide each disputed change. "Prepare a clean version" is an instruction about a file format, not authority to pick between two competing creative revisions. If a change is contested and nobody has been named to settle it, that change is not ready to be resolved. It is ready to be flagged.

Then save the authoritative draft untouched, under a name you will still recognize in three weeks, and make your edits on a copy. Two files, both named:

Slow Morning — treatment_v4 REVIEW.docx     (untouched, keeps every mark and comment)
Slow Morning — treatment_v4 CLEAN.docx      (the derivative you will send)

The review copy is not a spare. It is the only record of who proposed what, and the moment you accept and reject your way through a document, that record is gone from the working file. Keep it.

Put every reviewer back on the page

The revision decisions you are about to make should be made against a document that shows all of them. Set the display so tracked changes, formatting changes, and comment markers are all visible, and make sure every reviewer's marks are shown. Starting from No Markup is starting from a picture of the outcome, and the outcome is the thing you have not produced yet.

Here is a fictional one-page treatment to work through. It is invented for this article and has not been run in Word.

Slow Morning — 30-second treatment (fictional)

P1. A woman in a grey coat waits at a rain-slicked bus shelter. P2. On her counter at home, the Ashbury kettle's brushed steel body holds the last of the morning light. P3. She does not board. She walks back the way she came. P4. In the kitchen, the Ashbury trembles, steams, and clicks off. She pours.

Two reviewers have been in it: Maya Alarcón, the agency producer, and Devan Iyer, the creative director. Four items are present.

Three of these are decisions. The fourth is a question, and the difference matters more than anything else in this article.

The misleading view. Suppose you switch to No Markup. The page looks nearly settled: P2 seems to read without "brushed steel," the bold is in place, Maya's sentence sits in P3. It looks like the file where the work is done. But the deletion is still only a proposal, the comment is still attached, and nothing has been accepted or rejected. If you send that file, the revision state travels with it, and whether the recipient sees struck-through text and a comment marker depends on their view setting rather than yours. You have not made a clean file. You have made a clean-looking window.

The filtered view. Now suppose you filter the display so only Maya's changes show. Her one insertion remains; Devan's strike-through and comment marker vanish from the screen. The page looks very nearly clean. But Devan's two items did not go anywhere — they are simply not displayed. Microsoft's documentation notes that some controls act on the shown reviewers only and distinguishes "All" from "All Shown." So the visible set is not automatically the whole document, and a bulk command issued over a filtered view may not do what you think it does.

Why bulk commands are the wrong tool here. There is an "accept everything" command, and it would make the interface perfectly quiet. It would also accept R2 — the deletion the client rejected — and quietly produce the wrong sentence in the file the designer receives. With four items on one page, going through them one at a time costs you two minutes. The bulk command costs you the difference between the treatment that was approved and the treatment that was not.

So: find R1's insert, accept it. Find R2's struck-through words and reject that change — which restores the words rather than retyping them. Find R3's formatting change and accept it. Retyping "brushed steel" by hand would produce the right sentence and destroy the record that Devan cut it, which is exactly the record you will want if the client changes their mind on Thursday.

Comments decide nothing on their own

A comment is not a pending edit. Accepting and rejecting does not touch it, and deleting it does not answer the question inside it. Treat those as two separate jobs in a fixed order: settle the question, then remove the comment.

In the fixture, R4 asks whether the close stays as prose or becomes a shot list. That is a structural question about the treatment, and the person who owns it has not replied. You cannot accept it and you cannot reject it, because it is not a revision. You also cannot resolve it yourself, because the decision is not yours — that part is out of scope for a handoff. What you can do is keep the question alive in the review copy and keep the affected passage out of a file that will be treated as final.

That is the honest shape of this handoff. The clean derivative carries P1 through P3 in their decided form. P4 is withheld, and the note you send names it, says who owns the question, and says that the close is not final. A partial handoff that says so is more useful than a complete-looking one that will change. If the designer genuinely cannot start without the close, then the next move is to get the director's answer, not to guess at it in a file.

One related switch: turning Track Changes off stops new marks from being recorded. It does not remove the existing ones, and it is easy to mistake the two, which is roughly the confusion in a user question posted on Microsoft's Q&A site on 19 March 2024 — someone trying to hide or disable tracking and still ending up with a document or PDF that showed revision marks. That is one person's report rather than a documented cause, but the shape of it is familiar: the setting changed, the file did not. Leftover tracking is also worth checking before the next review round, so the next set of edits is not recorded on top of an audit trail nobody expected.

Reopen the file you are about to send

Your open window proves nothing. Save the derivative, close it, and reopen that exact file — the one whose name is in your delivery message. Then check it three ways.

Remaining state. Confirm no tracked revisions and no comments remain in the file you are sending. In the fixture, that means R1 has been accepted, R2 rejected, R3 accepted, R4's comment gone from the derivative and intact in the review master.

Wording and artifacts. Read the paragraphs rather than scanning the markup. Compare them against the decision sheet line by line: does P3 carry Maya's sentence, does P2 still read "brushed steel" with "Ashbury" bold, is P4 absent? Accepted insertions and rejected deletions can leave a double space, a missing space, or a bold run that has swallowed the punctuation after it. Catching those now is a one-minute fix; catching them after the designer has set the page is not.

The reading export. If a PDF is part of the deliverable, generate it from the clean derivative, not from the review master, and open the PDF to check it. A PDF is a rendering of the text at the moment you made it. It cannot resolve anything, and it will happily carry whatever state the source file had.

One more thing people reach for at this stage: Word's Document Inspector. It exists, but which items it offers and what it reports vary by version and platform, so treat its output as version-specific rather than as a guarantee that nothing personal remains in the file. If you use it, use it on the derivative. Running it across the review master is one way to destroy the record you deliberately kept.

What the note says

The handoff is finished when you can name everything in it. In the fixture, that is:

• The review master, untouched: Slow Morning — treatment_v4 REVIEW.docx, holding R1, R2, R3 and Devan's open comment on P4.
• The clean derivative you are sending: Slow Morning — treatment_v4 CLEAN.docx, carrying P1 through P3.
• The authority behind the wording: the client's written approval for R1 and R2, house style for R3.
• What is not in it and why: the close is withheld while the director decides whether it stays as prose or becomes a shot list.

That list takes four lines and prevents the most common failure in this whole operation — sending a file that looks finished to someone who will treat it as final. The quiet page is not the goal. The goal is a file in which you can account for every word, and a review copy that still remembers everything you decided not to decide yet.

Frequently asked questions

What does a genuinely clean tracked-changes handoff mean?

It does not mean a page that merely looks quiet. It means a saved copy in which every tracked change has been accepted or rejected on someone's authority, every comment has been handled on purpose, and you have reopened the exact file you intend to send to confirm both. Hiding markup changes what you see, not what the document contains. Microsoft's support documentation notes that hiding markup does not remove tracked changes, changes need explicit acceptance or rejection, comments are deleted through a separate operation, and some controls act only on the reviewers currently shown.

Why keep a review master and work on a copy?

The review copy is the only record of who proposed what, and accepting and rejecting your way through the working file destroys that record. Save the authoritative draft untouched under a recognizable name and make edits on a derivative. Before deciding, set the display so tracked changes, formatting changes, and comment markers are visible, and make sure every reviewer's marks are shown. Starting from No Markup shows a picture of the outcome before you have produced it.

Why are bulk accept commands risky in the invented Slow Morning fixture?

An 'accept everything' command would accept R2, the deletion the client rejected, and quietly produce the wrong sentence in the file the designer receives. With four items on one page, going through them one at a time costs little. Accept R1's insertion, reject R2's struck-through words to restore them, and accept R3's formatting change. Retyping the words by hand would produce the right sentence but destroy the record that Devan cut it.

How should comments be handled in a clean handoff?

A comment is not a pending edit. Accepting and rejecting does not touch it, and deleting it does not answer the question inside it. Treat those as separate jobs in a fixed order: settle the question, then remove the comment. If the owner has not replied, as with R4 asking whether the close stays prose or becomes a shot list, you cannot accept it, reject it, or resolve it yourself. Keep the question alive in the review copy and keep the affected passage out of the file that will be treated as final. A partial handoff that names the open question is more useful than a complete-looking one that will change.

What checks should happen after saving the clean derivative?

Close and reopen the exact file whose name is in your delivery message. Confirm no tracked revisions and no comments remain in the derivative: R1 accepted, R2 rejected, R3 accepted, and R4's comment gone from the derivative but intact in the review master. Read the paragraphs against the decision sheet rather than scanning markup: P3 carries Maya's sentence, P2 still reads 'brushed steel' with 'Ashbury' bold, and P4 is absent. Watch for spacing or bold-run artifacts left by accepted insertions and rejected deletions. If a PDF is part of the deliverable, generate it from the clean derivative, not the review master, and open the PDF to check it. Word's Document Inspector exists, but what it offers and reports varies by version and platform, so treat its output as version-specific; if you use it, use it on the derivative, not on the review master.

Further reading

• Merge Parallel Series-Bible Edits Without Losing Either Writer’s Changes — When two reviewers return separate marked files rather than one, this is the Combine route for preserving both writers' attributable edits without overwriting either return.
• Bring PDF Review Comments Back Into the InDesign Treatment — If the next review round arrives as PDF annotations instead of tracked changes, this shows how to bring those comments into the source and give each a disposition.