What a “commentary script” needs to do (and what it must avoid)
A commentary video isn’t just “having opinions on camera.” A strong commentary script is a guided argument that stays fair, stays readable, and earns trust. Your job is to help the viewer track: what the topic is, what you believe, why you believe it, what a smart opponent would say, and what your reaction means.
Common failure modes to design against:
- Rambling: jumping between points without a clear claim or destination.
- Straw-manning: attacking a weak version of the opposing view.
- Vibes-only: strong tone, weak evidence.
- Unclear stakes: viewers don’t know why they should care.
- Reaction without framing: “That’s crazy” with no explanation of what it implies.
The 6-part commentary template
1) Frame the topic and the stake (why this matters)
This is the “fair framing” moment: define what’s being discussed and why it matters beyond drama. Keep it specific and measurable when possible (money, safety, time, trust, incentives, misinformation, creator economy, consumer outcomes).
Fill-in:
- Topic: “Today we’re looking at [event/claim/trend].”
- Stake: “This matters because it affects [who] by changing [cost/risk/incentive/outcome].”
- Scope: “I’m focusing on [one angle], not [other angle].”
Example: “Today we’re looking at the claim that ‘AI thumbnails are killing creativity.’ This matters because thumbnails shape what gets watched, which shapes what gets funded. I’m focusing on whether AI tools change incentives for creators—not whether AI art is ‘good’ or ‘bad.’”
- Listen to the audio with the screen off.
- Earn a certificate upon completion.
- Over 5000 courses for you to explore!
Download the app
2) State the thesis early
Your thesis is your main claim in one sentence, stated before you start piling on examples. It should be debatable, narrow, and stable (it shouldn’t change every time you add a new clip).
Thesis patterns:
- “X is happening because Y, and the result is Z.”
- “The real issue isn’t X; it’s Y.”
- “X can be good, but only if Y; otherwise Z.”
Example: “My take: AI thumbnails don’t kill creativity; they raise the baseline, which forces creators to compete on clearer ideas and stronger packaging—unless platforms reward sameness.”
3) Present evidence and examples
Commentary still needs proof. Evidence can be: direct quotes, on-screen clips, data, timelines, policy text, side-by-side comparisons, or concrete first-hand observations. The key is to connect each piece of evidence to a claim.
Rule: one point = one main claim + the best support you have + what it means.
| Evidence type | What it looks like in a script | What it does |
|---|---|---|
| Quote/clip | “They said: …” then paraphrase | Anchors fairness; reduces misrepresentation |
| Example | “Here’s a case where…” | Makes abstract claims concrete |
| Comparison | “Before vs after” | Shows change, not just opinion |
| Numbers | “According to…” (keep it simple) | Prevents vibes-only arguments |
| Mechanism | “This happens because…” | Explains cause-and-effect |
4) Steelman the opposing view
Steelmanning means presenting the strongest reasonable version of the other side, then responding to that version. This increases credibility and prevents the video from feeling like a dunk-fest.
Steelmanning steps:
- Name the best argument: “The strongest version of their point is…”
- Grant what’s true: “They’re right that…”
- Set the boundary: “Where I disagree is…”
- Respond with a mechanism: “Because if we follow that logic, then…”
Example: “The strongest version of the ‘AI kills creativity’ argument is that when everyone can generate polished visuals instantly, originality gets drowned out. And they’re right that platforms often reward what looks familiar. Where I disagree is blaming the tool instead of the incentive structure—because the same pressure existed with templates, trend cycles, and thumbnail ‘rules.’”
5) Reaction beats (what surprised you, what it implies)
Reaction beats are where you show your human processing—without replacing reasoning. They work best when they translate emotion into meaning.
Two useful reaction formats:
- Surprise → explanation: “What surprised me is [detail]. That suggests [mechanism].”
- Contrast → implication: “They say [claim], but the evidence shows [contrast]. That implies [consequence].”
Example: “What surprised me is that the loudest complaints come from mid-sized creators, not beginners. That suggests the fear isn’t ‘I can’t make thumbnails’—it’s ‘my advantage is shrinking,’ which changes how competitive the niche feels.”
6) Closing with a clear takeaway
In commentary, the takeaway should be a crisp “so what” that the viewer can repeat. It can be a principle, a prediction, or a decision rule.
Takeaway patterns:
- Principle: “Don’t optimize for X; optimize for Y.”
- Prediction: “If X continues, expect Y.”
- Decision rule: “If you see X, ask Y.”
Structure map: “claim → support → implication” cycles
Most coherent commentary scripts are a chain of mini-arguments. Each mini-argument is a cycle:
Cycle #n (20–60 seconds each): Claim → Support → Implication (optional: Counter → Response)How to write each cycle:
- Claim: one sentence that can be disagreed with.
- Support: 1–3 pieces of proof (clip, quote, example, number, comparison).
- Implication: what this changes (incentives, outcomes, who benefits, what to watch for).
- Optional Counter: the best objection to your claim.
- Response: what your claim still explains better, or the condition where the counter is true.
Micro-template you can paste into a draft:
[CLAIM] The key issue is ________. [SUPPORT] We see this in ________ (example/clip), and ________ (second support). [IMPLICATION] That means ________ for ________. [COUNTER] A fair pushback is ________. [RESPONSE] That’s true when ________, but in this case ________.Checklist to avoid rambling (commentary edition)
- One thesis, one sentence: can you highlight it in your draft?
- Every paragraph has a job: label it in the margin as Frame, Claim, Evidence, Steelman, Reaction, or Takeaway.
- No “clip tourism”: don’t add a clip unless it proves a claim or changes the implication.
- Limit qualifiers: if you say “maybe,” “kind of,” “sort of,” replace with a condition: “This is true when…”
- Keep a running question: “What does this change?” If a sentence doesn’t answer it, cut or move it.
- One counterpoint per major claim: too many counters can dissolve your stance.
- Reaction must cash out: after any reaction line, add “which implies…”
- Define your target: are you critiquing a person’s claim, a system incentive, or an outcome? Don’t switch targets mid-point.
- End each cycle: finish with an implication sentence before starting the next claim.
Practical step-by-step: build a commentary segment using the template
Step 1: Write the frame in 3 lines
- Topic line
- Stake line
- Scope line
Step 2: Write the thesis in 1 line
Make it falsifiable: someone could reasonably say “no, that’s wrong.”
Step 3: Outline 3 claim cycles
For each cycle, write:
- Claim (1 sentence)
- Support (2 bullets)
- Implication (1 sentence)
Step 4: Add one steelman counterpoint
Pick the strongest objection to your overall thesis (not a minor nitpick). Add it after cycle 2 or before cycle 3.
Step 5: Add reaction beats as connectors
Place reaction beats at moments where the viewer might feel uncertainty or where a detail changes the meaning of the story.
Step 6: Draft the takeaway as a decision rule
Write it so a viewer could apply it tomorrow.
Practice: write a 3-point commentary script (with one counterpoint and one resolution)
Prompt: Choose a topic you can discuss fairly without needing private info. Example topic: “Creators should stop apologizing for sponsored segments.”
Fill-in worksheet (write your own)
- Frame (topic + stake + scope): ________
- Thesis: ________
- Point 1 (claim/support/implication): ________
- Point 2 (claim/support/implication): ________
- Counterpoint (steelman): ________
- Resolution (your response/condition): ________
- Point 3 (claim/support/implication): ________
- Takeaway (principle/prediction/decision rule): ________
Worked example (model script you can imitate)
Frame: Today we’re talking about creators apologizing for sponsors. This matters because sponsorships fund free content, but constant apologizing trains audiences to distrust the business model. I’m focusing on how the apology affects viewer perception—not whether any specific sponsor is “good.”
Thesis: Creators should stop default-apologizing for sponsors and instead frame sponsorships as a value trade with clear boundaries.
Point 1 — Claim: Apology language makes the sponsor feel like a betrayal, even when the ad is normal.
Support: When a creator says “Sorry, but we have to do this,” it signals the segment is low-quality and unwanted. Viewers learn that “sponsor” equals “time to skip,” which lowers performance and pressures creators to cram more ads elsewhere.
Implication: The apology doesn’t protect trust; it quietly erodes it by labeling your own monetization as shameful.
Reaction beat: What surprised me is how often the apology is automatic—like a verbal tic. That implies it’s not a strategic choice; it’s a habit you can replace.
Point 2 — Claim: A clear, confident sponsor frame can increase trust more than an apology does.
Support: If you say “This sponsor makes the video possible, and I only take deals that meet X,” you’re giving a standard the viewer can evaluate. If you add “The sponsor has no say in the opinion section,” you separate funding from conclusions.
Implication: You’re not asking for forgiveness; you’re offering transparency and boundaries, which is what trust actually needs.
Counterpoint (steelman): Some viewers have been burned by deceptive ads, so an apology can feel like empathy—acknowledging that ads are annoying and that the creator is on the viewer’s side.
Resolution: That empathy is valid, but you can express it without undermining yourself: swap “sorry” for “I’ll keep this short, it’s relevant, and then we’re back to the video.” Empathy plus structure beats shame plus vagueness.
Point 3 — Claim: The best long-term play is to treat sponsors as part of the product experience, not an interruption you hope people forgive.
Support: A sponsor segment with a clear promise (what the viewer gets), a boundary (what the sponsor doesn’t control), and a quick transition back feels professional. Over time, viewers stop seeing sponsorship as a moral compromise and start seeing it as the business model.
Implication: You reduce skip behavior, protect your credibility, and make it easier to say “no” to bad fits because your standard is public.
Takeaway (decision rule): If you’re about to say “sorry for the sponsor,” replace it with: “Here’s why this is here, here’s my boundary, here’s the value, and we’re back.”
