How to Revise a Treatment After Notes (Without Losing the Idea)

Key Takeaways

• A note is a symptom. Treat it as a diagnosis and you will fix the wrong thing correctly.
• Log every note verbatim before you interpret any of them. Interpreting as you read produces a defensive reading.
• Notes cluster. Six notes usually point at two problems, and solving the two is faster and better than addressing the six.
• Keep every draft. The most expensive failure in a multi-round revision is quietly reversing an earlier decision nobody remembers making.
• Some notes should not be actioned, and saying so with a reason is a normal part of the process rather than a confrontation.
• Reply with what changed and why before anyone reads the new draft, or the next round of notes will be about the same things.

Notes Are Evidence, Not Instructions

The person giving notes is an expert in their reaction and an amateur in your craft. Both halves matter.

Their reaction is real data. If a commissioner says the second act drags, it drags for them, and that is not up for debate. What they say to do about it is a guess, usually the first plausible fix that occurred to them, and it is frequently the worst available option because it treats the surface.

The classic pattern. "Can we cut the interview section, it feels slow." The interview section is not slow. It follows two minutes with no stakes established, so by the time the interview arrives the viewer has no reason to care. Cut the interview and the film loses its best material and still feels slow, because the actual problem was upstream. Establish the stakes earlier and the same interview plays fast.

That is the entire discipline. Believe the symptom completely. Distrust the prescription.

The exception. Some notes are instructions and should be taken literally: legal, factual accuracy, brand rules, budget, contractual requirements. Do not go looking for the deeper meaning in "we cannot say the competitor's name". Sort these out first, because they are not creative decisions and treating them as such wastes everyone's time.

The Method

Step 1: Log everything before you react

Write every note down verbatim, attributed to whoever gave it. Do not paraphrase, because paraphrasing is where interpretation sneaks in.

If notes arrived by phone or in a meeting, write them up the same day, and send them back to the person to confirm you heard correctly. This one habit removes a large share of later disputes, since it converts "that is not what I said" into a checkable claim.

Read the whole set once without deciding anything. The first read is for absorbing, not for planning, and the plan you make while defensively reading the first note will not survive the fifth.

Step 2: Translate each note into a problem

For each note write the underlying diagnosis in your own words.

Some notes will resist translation. Write "unclear, need to ask" and ask. A note you cannot diagnose is a note you will address wrongly.

Step 3: Cluster the problems

Now group the diagnoses. Eleven notes routinely reduce to three problems.

This step is where a revision becomes tractable. Eleven separate fixes produce a patchwork document that reads as though it were assembled by committee, because it was. Three structural fixes produce a coherent draft.

Clustering also reveals disagreement in the room. When two notes translate to contradictory problems, you have found something more important than either note: the stakeholders do not agree, and no draft will satisfy both. Surface it and ask, rather than splitting the difference, which produces a draft neither of them wanted.

Step 4: Decide what you are not doing

Not every note gets actioned. Deciding this explicitly, in writing, is what keeps a treatment coherent across four rounds.

Three legitimate reasons to decline: the note conflicts with a decision the client already approved, the note conflicts with another note and you need a ruling, or the note would break something load-bearing that the note-giver cannot see.

That third one obliges you to explain the load. "The slow open is doing work: it is what makes the reveal at four minutes land. If we speed it up we lose the reveal. I would rather cut ninety seconds from act two." That is a professional response and it is usually accepted. Silently ignoring the note is not, and it is what most people do.

Step 5: Revise against the problems

Now write, addressing the clustered problems rather than the individual notes.

Work top down. Structure first, then sequence, then scene-level, then line. A revision that starts with line edits is polishing a structure you are about to change.

Step 6: Keep the previous draft visible

This is the step that fails silently and expensively.

By round four, a change made to satisfy a new note frequently reverses a decision made in round one for a reason nobody now remembers. The client sees the reversal, cannot articulate what is wrong, and gives you a note that is really a request to go back to something they liked two rounds ago.

The defence is keeping the drafts and their notes side by side, so that before making a change you can see whether the thing you are about to remove was put there deliberately.

Version history in a document handles part of this and is hard to read across four rounds, because it shows you what changed without showing you why. On a Storyflow canvas each draft sits as an object with its notes and the one-line reason for each change attached, so the whole revision history is visible at once rather than buried in a changelog. Its AI reads the whole board, which means you can ask what changed between two drafts and why, and get an answer grounded in your actual notes.

Step 7: Reply before they read

Send a short summary with the new draft: the problems you identified, what you changed, and what you did not change with the reason.

Six or eight lines. It is the highest-return part of the process, because a reader who knows what you were solving reads the draft looking for the solution. A reader who receives a draft with no context reads it cold and re-notices everything they noticed last time.

Round Discipline

Rounds should get smaller. If round three produces more notes than round two, something is wrong at a level the notes are not addressing, and another revision will not fix it. Stop and have the conversation about what the piece is.

Agree the number of rounds up front. Two rounds included, further rounds chargeable, is standard and it changes behaviour: clients who know they have two rounds give better notes, because they consolidate rather than reacting.

Consolidate the note-givers. Notes from five people arriving separately will contradict each other and you will be the one discovering it. Ask for one consolidated set from a named person who has already resolved the internal disagreements. This is the single most effective thing you can negotiate.

What Goes Wrong

Every note actioned literally. Produces an incoherent draft that satisfies each note and nobody. The commonest failure, and it comes from treating notes as a checklist.

Notes actioned in the order received. The order reflects the order the reader noticed things, which correlates with nothing structural.

Disagreement split down the middle. Two contradictory notes both half-addressed produces a draft that fails twice. Escalate instead.

The previous draft is gone. Round four undoes round one. Nobody can explain why the piece got worse.

No reply sent. Next round's notes repeat this round's, because the client had no way to know what you were solving.

Declined notes not mentioned. The client assumes you missed them and repeats them with more force.

The Bottom Line

Believe the symptom, distrust the prescription. Log the notes verbatim, translate each into the problem underneath, cluster the problems, decide explicitly what you are not doing and why, then revise structure first.

Keep every draft with its notes attached so round four cannot quietly undo round one, and always send the reasoning before the draft. That is what turns a revision from damage control into an actual improvement.