Track changes tools—Microsoft Word’s Track Changes, Google Docs’ Suggesting mode, or comment/diff conventions in LaTeX editors like Overleaf—are how most co-authored manuscripts actually get written today. The tools themselves are simple. What causes friction is the absence of shared norms for using them: silently accepting a colleague’s edit, leaving a comment thread unresolved for weeks, or submitting a manuscript that still has another author’s name attached to every tracked edit in the file metadata. This guide covers the etiquette—the working agreements a co-author team should set explicitly, ideally before drafting starts—not the mechanics of who holds the master copy, which is covered in CASRAI’s guide on managing manuscript versions across co-authors.
Why this is a distinct problem from version control
Version control answers “which file is current.” Track-changes etiquette answers a narrower question that persists even once everyone agrees on the current file: how should co-authors mark, explain, and resolve the edits inside that file so that the team’s collective intent is legible, no one’s contribution is silently overwritten, and the final submitted version is one every listed author can honestly say they approved. That last point is not just good practice—the International Committee of Medical Journal Editors (ICMJE) authorship criteria require “final approval of the version to be published” from every author, alongside drafting or substantively revising the work and being accountable for its accuracy. Approval of a specific, final version is hard to give honestly on a manuscript where tracked edits were never resolved, comments were never addressed, and no one can say for certain what the “final” text actually is.
Set the working agreement before drafting starts, not after conflict appears
1. Agree on one tool and one mode
Pick a single platform for the drafting phase—Word with Track Changes, Google Docs in Suggesting mode, or a LaTeX workflow with visible diffs—and commit the whole author team to it. Mixed workflows (one author edits a Word file with Track Changes off, another pastes revisions into a separate Google Doc) are the single most common cause of silently lost edits, because there is no longer one authoritative change history to reconcile against.
2. Never accept or reject another author’s tracked change unilaterally
Accepting a change removes the record that a suggestion was made and by whom; rejecting one removes it entirely. Either action, taken without at least a comment explaining the reasoning, erases a co-author’s input without their knowledge until they open the file again—by which point the disagreement is harder to resolve because the original edit is gone. The safer default: leave substantive edits as open suggestions until the author who proposed them, or the corresponding author acting as tie-breaker, has had a chance to see the response.
3. Use comments to explain non-obvious edits, not just to make them
A tracked deletion or rewording is legible on its own; the reasoning behind it usually is not. A comment attached to a substantive change (cutting a sentence, altering a claim, reordering a paragraph) turns a silent edit into a proposal the rest of the team can evaluate quickly, rather than a change someone has to reverse-engineer the intent of before agreeing or objecting.
4. Distinguish copyedits from substantive edits
Not every tracked change carries the same weight. Fixing a typo, a citation format, or a unit abbreviation is different from altering a claim in the Discussion section or restructuring the Methods. Some teams handle this with a convention as simple as flagging substantive edits with a comment or a specific tag (e.g., “[SUBSTANTIVE]”), so co-authors reviewing a long diff can prioritize the changes that actually need their sign-off rather than reading every comma fix with the same scrutiny.
5. Resolve comment threads deliberately, don’t just delete them
Both Word and Google Docs let a comment be marked resolved (or deleted) once it’s addressed. Deleting a comment thread without resolving the underlying disagreement leaves no record that the point was ever raised—useful in the moment, costly if the same disagreement resurfaces during peer review and no one remembers it was already discussed once. Marking a thread resolved, rather than deleting it, preserves that history while still clearing it from the active view.
6. Set an explicit turnaround expectation
A shared document with no agreed review window tends to stall on whoever is slowest to respond, and it is often unclear whether silence means agreement or means the co-author hasn’t looked yet. Stating a turnaround window when a draft is circulated (e.g., comments by a specific date, or an explicit note that no response will be treated as no objection) avoids both a manuscript stalling indefinitely and a co-author later objecting to a change they never actually reviewed.
7. Produce one clean, fully-resolved version before submission
Before a manuscript goes to a journal, every tracked change should be either accepted or rejected (nothing left as an open, unresolved suggestion) and every comment thread closed. A submission with visible tracked changes or open comments is not just an appearance problem—editorial systems generally expect a clean, final manuscript, and a stray tracked edit or an unresolved internal comment (occasionally including candid disagreement between co-authors) reaching an editor or reviewer is avoidable and unprofessional.
Strip identifying track-changes metadata before blind submission
Journals that run single- or double-blind peer review expect a manuscript free of author-identifying information, and tracked-changes metadata is an easy thing to overlook. Word embeds each editor’s name (from the document’s author/user settings) against every tracked change and comment, which survives in the file even after changes are accepted unless it is explicitly removed—Word’s built-in Document Inspector (under File > Info > Check for Issues) can find and strip this along with other hidden metadata. Google Docs carries similar per-suggestion attribution; the reliable way to clear it for a blind submission is exporting a clean copy (File > Download, or File > Make a copy after accepting all suggestions) rather than assuming a visually clean document has no residual metadata. This is a genuinely easy step to miss on a deadline, and it is worth a specific checklist item before any blind submission, not an assumption that “accept all changes” alone handled it.
Handling disagreement over an edit
Disagreement about a specific change is normal on any multi-author manuscript and is not itself a breach of etiquette—how it is handled is what matters. A few working patterns:
- Default to the corresponding author as tie-breaker for disagreements that don’t resolve after brief discussion in comments, consistent with that person’s role coordinating the manuscript through submission (see CASRAI’s guide on corresponding author responsibilities under ICMJE).
- Defer to domain expertise on substantive scientific edits—the co-author whose contribution the CRediT taxonomy would classify under a given role (e.g., Formal analysis, Methodology) generally has the strongest claim to the final call within their area, though final approval still rests with the full author group per ICMJE’s criteria.
- Take unresolved disagreement out of the comment thread and into a direct conversation once a thread has gone back and forth more than two or three times—a live discussion resolves ambiguity a written thread often just prolongs.
Tool-by-tool notes
Microsoft Word (Track Changes)
The most established option, with per-author color-coding, a Reviewing Pane that lists every change in sequence, and the ability to accept/reject changes individually or in bulk. Its main etiquette risk is the identifying-metadata issue above, and the fact that Track Changes can be silently turned off by any editor with the file open—worth explicitly confirming it’s enabled at the start of each editing session on a shared file.
Google Docs (Suggesting mode)
Real-time collaborative editing means co-authors can see each other’s suggestions as they’re made, which speeds up back-and-forth but also means a colleague can be mid-edit on the same paragraph simultaneously. Suggesting mode must be deliberately selected (rather than Editing mode, which makes direct, untracked changes)—worth restating as a team norm, since it’s easy for a collaborator to forget which mode they’re in.
LaTeX / Overleaf
LaTeX has no native equivalent to Word’s Track Changes; teams typically rely on Overleaf’s built-in track-changes feature, the latexdiff command-line tool to generate a marked-up PDF comparing two versions, or simply disciplined use of % comments and version control (Git) commit messages to document who changed what and why. Because none of these are as visually integrated as Word or Google Docs, LaTeX teams benefit even more from an explicit convention—agreed at the outset—for how a proposed change gets flagged, discussed, and accepted into the shared source file.
Frequently asked questions
Is it acceptable to accept my own tracked changes but leave a co-author’s open?
Yes—accepting your own edits doesn’t remove anyone else’s input, since you authored the change yourself. The etiquette concern is specifically about accepting or rejecting someone else’s suggestion without their input, not about clearing your own.
Who should have final say when co-authors disagree on a tracked edit?
There is no formal rule; in practice the corresponding author often serves as tie-breaker given their role coordinating the manuscript, though disagreements over subject-matter substance are better resolved by deferring to whichever co-author holds the relevant expertise or CRediT role. What matters more than who decides is that the decision is visible to the full author team before submission, consistent with ICMJE’s requirement that every author give final approval to the actual submitted version.
Does track-changes etiquette matter for author order or credit disputes?
Not directly—author order and contribution credit are governed by the team’s authorship agreement and, where used, a CRediT contributor-roles statement, not by who made the most edits in a shared document. See CASRAI’s CRediT guides for that separate question.
How do I make sure my manuscript is actually clean before submitting, not just visually clean?
Run the document through a metadata check—Word’s Document Inspector, or a fresh export from Google Docs after all suggestions are accepted—rather than relying on a visual scan, since accepted changes and closed comments can still leave author-identifying metadata embedded in the file.
