rival-breakdown

Rival Breakdown

This skill studies the videos your competitors already made on a topic and plans a better one for you. Paste the transcripts of two or more competitor videos (YouTube, podcast episodes, webinars). It breaks down each video (who it talks to, how it hooks, what belief it installs, what value it delivers, where it falls short), lines them up side by side, finds the gap none of them fill, and hands back an outline for a video built to own that gap. With one video it gives the breakdown only. Use it before you script any video on a topic other creators have already covered, when a competitor's video took off and you want to know why, or when you want to find the gap before you plan your own.

Other skills in this pack cover nearby jobs. To turn the outline into a script you can record, use script-writer. For the title and thumbnail, use title-and-thumbnail. For a wider competitor scan beyond video, use market-researcher.

Before you start

1. Read the Business Context file first. The Business Context file is a short doc in your Claude Project: what you sell, to whom, your proof, your voice and the next step you want viewers to take. Look for it in the Project knowledge or in this chat. Pull out: what you sell, who buys it (their stage, their fears, what they value), how you want to be seen and what you refuse to be, your proof (client results, numbers, years in the field), your brand voice, and the next step you want viewers to take (for example, book a call). Never ask for something the file or the chat already answers.

2. Ask only for what is missing, in one short numbered message, four questions at most. Skip any question the file or the chat already answers. If nothing is missing, skip the questions and start. Priority order:

   1. If there is no Business Context file and the chat does not say: "What do you sell, to whom, and what should a viewer do after watching (book a call, grab a resource)?"
   2. The competitor videos: "Paste each transcript (two to five videos on the same topic), with the creator, the title and the length, and the views and upload date if you know them."
   3. Your credibility, if the file does not say: "Your proof in two lines: results, years in the field, numbers." The "Who you are" part of the outline needs it. If the user skips it, use [placeholder: your result].
   4. Optional: "Your topic or working angle, if you have one, and who this video is for if it is not your usual buyer."

3. Get full transcripts. Transcripts come first. Links work only if you can open the full spoken text.

   • On YouTube: open the video, click "...more" under it, then "Show transcript", and copy the text. If there is no "Show transcript" option, the video has no captions. Ask for a different video.
   • On other platforms: use the captions, the platform's transcript, or any transcript tool the user trusts.
   • Links: if web search or web fetch is on, you may try to open a link. A video page usually does not include the spoken text, and third-party transcript sites are often partial. Never analyze from the title, the description or the comments.
   • Completeness test: count a transcript as full only if it starts at the first spoken line and its word count is not far below the runtime at roughly 130 to 160 words a minute (a 10 minute video is about 1,300 to 1,600 words; fast talkers run higher). If the transcript has timestamps, the last one gives the runtime. If text you read from a link fails this test, ask the user to paste the transcript. If a pasted transcript fails it, say so and lower your confidence in that video's read.

4. No videos yet? Help the user pick them, in this order:

   1. Search YouTube for the exact phrase your buyer would type about this topic.
   2. Open three to five channels in your niche, go to each channel's Videos tab and sort by Popular.
   3. Pick two to five videos on the exact topic, from roughly the last 12 months, with views well above that channel's usual. Study winners, not random uploads. A video that flopped teaches less.
   4. Paste the transcripts.

   If web search is on, you may suggest candidate videos with their links for the user to confirm. You still need the full transcripts before the study starts.

How it works

Work through these steps in order. Write the result in the format under "What you get", not as a list of labeled steps.

Step 1. Pick the mode and the format

• One video: the breakdown only (Step 3 to Step 5). One video shows one creator's choices, not the market's gap, so do not build an outline. End with the gaps that look promising and ask for one or two more videos. If the user insists on an outline, build it and label it "based on one video".
• Two or more videos: the breakdown for each video, then the comparison, then the outline.
• More than five: ask which five matter most, or pick the five closest to the exact topic (favoring any with sourced view numbers well above their channel's usual) and say which ones you picked and why.

This skill is built for long-form videos (about 5 minutes or more), podcasts and webinars.

• Rival videos under 90 seconds (Reels, Shorts, TikToks, short LinkedIn videos): say so up front. Shorten each breakdown to four parts: the hook, the one value point, the main pattern and the ask. Still rate the awareness levels for Part 2. Replace Part 3 with the short-form outline under "What you get".
• Between 90 seconds and 5 minutes: use the full breakdown, and give the outline two or three core sections instead of three to five.
• A mix of short and long: study them in two groups, or ask which format the user plans to make.

Step 2. Set the lens

Write a short buyer profile from the Business Context file or the user's answers: who they are, where they are stuck, what they fear, what they value. Judge every video through this buyer, not through a general viewer. A video can be great for its own audience and useless for yours. That mismatch is often the gap.

Rate every video on these five awareness levels (an idea from copywriter Eugene Schwartz):

Level	Where the viewer is
1. Unaware	Does not see a problem yet.
2. Problem aware	Feels the pain, does not know a fix exists.
3. Solution aware	Knows fixes exist, unclear which one or how it works.
4. Product aware	Knows the options, including offers like yours, and is comparing.
5. Most aware	Trusts the offer, needs de-risking and proof to act.

Step 3. Read each video twice

First pass, the flow: where the hook ends, where each value point starts, where the call to action sits. Second pass, the mechanics: how each part moves the viewer's beliefs and feelings. Use timestamps when the transcript has them.

Step 4. Break down each video

For every video, work out these eight parts:

1. Awareness shift. The level the viewer starts at, the level they end at (a range is fine, like 3 to 4), and how their state changes along the way.
2. Hook. The main emotional lever (fear of loss, desire, curiosity, status, relief), the problem or desire it triggers, and its effect on how the viewer thinks and feels. Note how long the video takes to deliver its first real value.
3. Core thesis. One sentence: the belief the video installs in the viewer's mind.
4. Value structure. The narrative type (list, step by step, story, hybrid, other) and the value points in order.
5. Value mapping. For each value point: what is explained, why it matters to your buyer, and how it reduces confusion, risk or resistance.
6. Patterns used. The persuasion moves, each with a short piece of evidence from the transcript. Look for authority signals, simplifying (removing complexity), risk reversal, social proof, reframing, fear normalizing ("this is common, it is not your fault"), control restoring (giving the viewer a clear way to act), open loops, and any others you see.
7. Total perceived value. Cognitive: what the viewer understands better. Emotional: what feels safer, clearer or more believable. Behavioral: what mindset or action it nudges.
8. Strategic limits. Where it falls short for your buyer, and what angle it misses or avoids. Use the gap check in Step 5.

Step 5. Run the gap check

Test each video against these gap types. Name only the ones that clearly apply, each with evidence.

• Wrong level. Starts too high or too low for your buyer's awareness.
• What without how. Names the fix but never shows the method, the steps or the tradeoffs.
• Claims without proof. No results, examples, numbers or demonstrations.
• Dodged objection. Skips the question your buyer actually asks: cost, time, risk, "will this work for someone like me?"
• Wrong viewer. Talks to beginners, or to a different business size or model than your buyer.
• Fear without control. Raises fear or urgency, then leaves the viewer with no clear way to act.
• Hype. Big promises, buzzwords, "secret" framing. A skeptical buyer tunes out.
• Thin or stale. Old examples, missing steps, surface-level advice.
• No filter. Never says who this is not for, so the right viewer never feels singled out.
• Weak ask. The call to action asks too much too soon, or asks for nothing.

Step 6. Compare the videos (two or more)

Fill the head-to-head table (see "What you get"). Then sort everything the videos said into three groups:

• Table stakes: points most competitors cover. The viewer expects them. Your video covers them too, but faster.
• Over-indexed: claims, words or tactics every competitor leans on until they all sound the same. These feed the "Do not say" list.
• Gaps: what your buyer needs that no competitor covers well.

Then choose ONE angle to own. It must pass three tests: it is true for your business (check the Business Context file or the user's answers), your buyer wants it, and no competitor already owns it. If no angle passes all three, say so and name the closest option.

Step 7. Build the outline for your better video

Build it around the angle from Step 6.

• Hook: the context (what the video is about), the common belief (what most people think, and why it is partly true), and the reframe (your angle, which shifts the frame).
• Setup: three or four things the viewer will learn, and the cognitive outcome: which decision or belief changes by the end.
• Who you are: two or three lines from the Business Context file or the user's answers. Proof and real-world results, not theory. If proof is missing, leave [placeholder: your result].
• Core sections, three to five. For each: a title, the psychological objective, the key concepts, an example or analogy, a buyer filtering moment where it fits ("this is not for you if..."), and an open loop into the next section.
• Conclusion: the awareness shift in one line, a final rule or principle, a soft call to action that matches your next step (no pressure), and an optional tease of the next video.
• Tone and strategy notes: the target entry awareness level, the state you want the viewer in at the end, and the things NOT to say, based on what competitors over-indexed on.

Outline rules: table stakes go early and short. Most of the time goes to the gap. The angle shows up in the hook, in the core thesis and in at least two core sections. Formats (list, step by step, story) are fair game. Never copy a competitor's specific order of points, examples, stories, title or catchphrases.

Step 8. Check before you hand it back

• Every claim about a competitor video is backed by a short quote or a timestamp.
• Every video has an entry and an exit awareness level, each with a reason.
• Each core thesis is one sentence.
• The angle is backed by the Business Context file or the user's answers, not invented.
• Formats are reused freely, but no competitor's specific order of points, examples, stories, title or catchphrases is copied.
• Every view number has a source and a date, or says "not given". No made-up views, results or client stories.
• Any partial or failed transcript is flagged.

What you get

Always use this structure, written as normal formatted text (not inside a code block). Repeat Part 1 for each video. Parts 2 and 3 appear only with two or more videos. Keep each video breakdown to about 300 to 500 words. Never pad.

Rival breakdown: [topic]

Part 1. Video by video

[Creator] | [Video title] | [length]

Sign of success	[views and upload date, with the source and the date you read it; compared with the channel's usual only if sourced; otherwise "not given"]
Transcript	[full, or partial and why]
Awareness shift	[level] to [level]. [one line on how the viewer's state changes]
Core thesis	[one sentence]
Narrative type	[list, step by step, story, hybrid, other]
Primary lever	[emotional lever, and the problem or desire it triggers]

• Hook: [what happens in the opening, its effect, time to first real value]
• Value mapping:

#	Value point	Why it matters to your buyer	Confusion, risk or resistance it removes
1	[point]	[why]	[what it removes]

• Patterns used: [pattern]: [short quote or timestamp]
• Total perceived value: Cognitive: [...]. Emotional: [...]. Behavioral: [...].
• Strategic limits: [gap type]: [how it shows up, with evidence]

Part 2. Head to head

Creator	Awareness level	Positioning	Primary lever	Strategic limits	Where you win
[name]	[entry to exit]	[how they present themselves]	[lever]	[top gaps]	[clarity, proof, method, fit to your buyer]

• Table stakes: [points to cover, fast]
• Over-indexed: [what everyone leans on]
• Gaps: [what no one covers well]
• The angle to own: [two or three sentences: the key gaps, the angle, and why it passes the three tests]

Part 3. Your better video: "[working title]"

• Core thesis: [the one belief your video installs]
• Hook: Context: [...]. Common belief: [...], partly true because [...]. Reframe: [...].
• Setup: You will learn: [three or four points]. By the end: [the decision or belief that changes].
• Who you are: [two or three lines of proof from the Business Context file or the user's answers]

Section [number]: [title]

• Objective: [what should shift in the viewer]
• Key concepts: [...]
• Example or analogy: [...]
• Filter moment: [who this is not for, if it fits here]
• Open loop: [the question that pulls the viewer into the next section]

Conclusion

• Shift: [from where the viewer started to where they end]
• Final rule: [one line]
• Soft call to action: [matches your next step, no pressure]
• Next video tease: [optional]

Tone and strategy notes

• Target entry level: [level and why]
• End state: [how the viewer should think and feel]
• Do not say: [phrases, claims and tactics the competitors over-indexed on]

With one video, end Part 1 with Gaps worth testing (two or three bullets) and ask for one or two more videos to complete the study.

Short-form version (rival videos under 90 seconds). Part 1 becomes one table with a row per video:

Creator and title	Hook line	The one value point	Main pattern	The ask
[name, title, length]	[first spoken line, short quote]	[...]	[pattern with evidence]	[what it asks the viewer to do]

Part 2 stays the same. Part 3 becomes:

Part 3. Your short: "[working title]"

• Hook line: [the first spoken line, about 12 words or fewer]
• The one point: [the single idea, built on the gap]
• Proof: [one result or example from the Business Context file or the user's answers, or [placeholder: your result]]
• Call to action: [one small step that matches your next step]
• Target length: [seconds]

Short example of a "Strategic limits" line (placeholder: a sales coach studying a competitor's video on closing more calls):

• What without how: the video says "qualify harder" four times but never shows a single question to ask (around 6:10).
• Wrong viewer: every example is a first-year solo seller; your buyer runs a team of five closers.

Rules

• Diagnose, do not praise. Analytical, neutral, precise. Say what the video does and why it works or fails. No compliments, no "great video".
• Evidence or it does not count. Every claim about a competitor video points to a short quote (under 15 words) or a timestamp. Never guess at content you could not read.
• Out-position, never copy. Formats (list, step by step, story) are fair game. Never copy a competitor's specific order of points, examples, stories, title or catchphrases. Win through clarity, a sharper fit to your buyer, better proof and a real method.
• Never reproduce the transcripts. Summarize in your own words. Short quotes only, as evidence.
• Your buyer is the judge. A limit only counts as a gap if it matters to the buyer in the Business Context file or the user's answers.
• View numbers need a source. Every view count or upload date needs a source and a date (the user, or the page you read). Never estimate a channel's usual views from fewer than about 10 recent uploads. Never call a video an outlier ("3x the channel average") without that. Otherwise write "not given".
• Never invent proof or data. No made-up views, results, client names or testimonials. Use only what the user or the Business Context file provides. Otherwise leave a clear [placeholder].
• One angle, one outline. Commit. Not a menu of ideas.
• No hype. No emojis and no hype words in the analysis or the outline (unlock, game-changing, secret, skyrocket), unless your buyer uses them.
• Flag weak input. If a transcript is partial, auto-generated with errors, or missing its opening, say so and lower your confidence in that video's read.
• Critique the video, not the person. No mocking, no guesses about a creator's motives.
• Plain words. A busy founder should be able to read the whole study in one sitting.

Part of the free 20-skill pack by Ruben Davoli at BeaverMind (beavermind.ai).