How to Cut a Script from the Graph (A Cut List)

If you have one retention graph and one transcript, you already have enough to make a useful cut list. Start with a timestamp, map it to the exact spoken line, then choose CUT, REWRITE, or MOVE. Do not cut a spike because it looks unusual, and do not treat every dip as a verdict. The graph tells you where to inspect; the script tells you what to change.

What How to Cut a Script from the Graph (A Cut List) actually is (and what it is not)

A cut list is a timestamped edit decision list: a small table that joins a moment on your retention graph to the exact transcript line and the action you will take. The minimum useful row has five parts: time, line, evidence, diagnosis, and action. “The middle feels slow” is a feeling; “00:18, line 14, viewers leave during a repeated setup, CUT” is a testable edit note.

The graph is useful because it gives your review a location. It does not explain the sentence, the visual, the promise, or the joke at that location. Your transcript supplies that missing context. When you line up the two, you can decide whether the problem is excess wording, a late payoff, a misplaced explanation, or a moment that should stay because it is doing important work.

YouTube’s SOURCE 1 page is currently a Reach page. It supports the Analytics route and the relationship between impressions, views, and watch time; retention-specific guidance below is separately sourced or labelled as method.

Treat the shape as a locator, not a command. A dip says inspect here; a spike says understand what worked.

Why this shows up in YouTube Studio

YouTube’s current performance documentation says the Content tab gives a channel-level overview of what your audience watches and how it interacts with your content. It also lists key moments for audience retention under video analytics and says typical retention can compare your 10 latest videos of similar length. That is enough context to use the graph as a comparison and diagnosis tool without inventing a universal “good” number.

The current Studio workflow is a rule of thumb: open Analytics, select the video-level retention report, and find the timestamp you want to inspect. YouTube’s SOURCE 2 page says the retention report shows how different moments held viewers’ attention and that expanded reports can help you compare performance and export data. Record the date, report name, and selected video beside your screenshot.

Do not confuse a graph location with a cause. Read the sentence, visual beat, topic change, and promise around it before choosing an edit.

Retention data may take time to process. Use a settled report, or label an early cut list provisional and update it if the graph changes.

Worked example 1: the failure

Suppose an illustrative 8-minute upload has a dip at 00:18, a spike at 00:42, and another dip at 02:10. The transcript shows a repeated definition at 00:18, a concrete example at 00:42, and a long aside at 02:10. These timestamps and viewer counts are illustrative, not YouTube data; the point is the mapping method.

The failure is to see the first dip and cut the next 20 seconds without reading the script. That can delete setup while leaving the repeated definition. Cutting a spike makes the same mistake in reverse.

The bad list uses the graph as a red pen and records only “bad section.” For a joke or example, read setup and payoff separately so you do not cut the part that earned attention.

Worked example 2: the fix

Use the same illustrative upload, but rewrite the review sheet. At 00:18, mark REWRITE and quote the repeated line. At 00:42, mark KEEP and note that the concrete example created the spike. At 02:10, mark MOVE and place the aside after the main promise has been delivered. Every row now has a bounded action and a reason another editor can inspect.

Do not chase a perfect curve. Make the edit smaller and explainable: rewrite wording, move order, or cut repeated, delayed, or unrelated material.

For the next upload, shorten the opening definition, bring the example earlier, or move the aside after the first payoff. Change one scripting variable at a time and version the cut list.

How to check this in YouTube Studio (step by step)

1. Save the transcript. Keep the exact words, line breaks, and timestamps used in the edit. Save both versions if the transcript changed.

2. Choose the video and report. Open YouTube Studio, then choose the current video-level Audience retention report. The exact tab names can move; use the report that shows the video’s retention shape and record the date.

3. Wait for a usable report. Retention data can take 1–2 days to process. If you work earlier, label the list provisional and revisit it once the report has settled.

4. Capture the graph. Export or screenshot the view with 00:00, the reviewed moment, and its surrounding shape visible.

5. Mark moments, not every pixel. Start with meaningful dips, spikes, or changes of shape. A dense list of every tiny movement creates activity without a decision.

6. Map time to lines. Find the exact sentence, visual beat, and transition at the timestamp. If the graph time and transcript time do not align, document the offset instead of forcing a false match.

7. Write one diagnosis. Use concrete language such as repeated wording, late payoff, topic jump, slow setup, or misplaced detail.

8. Choose one action. Mark CUT, REWRITE, or MOVE. You may add KEEP for a spike or a strong section, but do not hide the required action labels under vague notes like “fix pacing.”

9. Review the surrounding shape. Check several seconds before and after the moment. An isolated movement can mean something different from a long decline or a recovery spike.

10. Export the handoff. Give the editor the graph, transcript version, cut list, and one-sentence goal. For a structured second pass, use the YouTube script review tool exactly once in the workflow; do not let the tool replace the timestamp-to-line evidence.

For script structure, compare your decision with How to Script a YouTube Video. If the problem is order rather than wording, use YouTube Script Outline Template before rewriting the whole draft.

The trap

The trap is treating the retention graph like an automatic editing command. A red-looking dip feels decisive, so you cut the nearest lines. A tall spike feels like noise, so you smooth it away. Both reactions skip the transcript, which is where the actual edit decision lives.

A joke is a good stress test. If the dip happens during the setup and the spike happens at the payoff, cutting the whole beat would remove the evidence that the idea worked. Shorten the setup, move the joke after the promise, or keep it if the surrounding context supports it. Name the choice and leave the graph screenshot attached.

If you want to shorten the review loop, RetentionYT can help organize repeated retention decisions, but the manual method works without signing up. The product should make a clear method faster, not turn an unclear graph into an automatic edit.

What to do in the next upload

Before you publish, save the transcript version you will use for the post-upload review. After the retention report is usable, make a small cut list instead of a full rewrite. Keep the graph screenshot beside the row so someone else can inspect the same moment.

A cut list is successful when another editor can reproduce your decision. They should know which graph moment you saw, which line you read, what you think happened, and what action you recommend. If the row cannot answer those four questions, it is still a hunch.

Use the graph to find the work. Use the transcript to define the work. Then decide whether the fix belongs in the published edit or the next upload. That is how a retention curve becomes a script tool rather than a dashboard you stare at after the fact.

Source note: Product and Studio claims are limited to the two locked YouTube Help sources opened on 24 August 2026; the cut-list labels and examples are editorial method, and all numerical examples are illustrative.