Act as a practical specialist helping me with youtube thumbnail testing system. My context: [channel niche, current CTR, design tools, upload frequency].
1. Design 3 thumbnail concepts per video.
2. Test faces, text, and contrast variations.
3. Change underperforming thumbnails after 48 hours.
4. Keep the swipe file of what works.
Use my details, not a generic example. If a fact needed for the plan is missing, ask for it or mark the assumption. Keep the result ready to use.
#creators#thumbnails
💡 Replace the [brackets] with your details — the more specific, the sharper the answer.
The first 30 seconds engineered so nobody clicks away.
Act as a YouTube scriptwriter whose videos hold 50%+ retention. My video topic: [topic]. Video length target: [minutes]. My channel's style: [describe or paste a past intro]. My audience: [who they are, what they clicked hoping to get].
Write my hook:
1. The cold open: 3 versions, each 15-30 seconds spoken. Version A opens with the end result shown first. Version B opens mid-story at the tension peak. Version C opens with the bold claim the video will prove.
2. For each: the exact spoken words, the visual on screen, and the 'open loop' it plants — the question the viewer now needs answered.
3. The title-thumbnail promise check: the hook must pay off the click, not extend the mystery past patience. Check each version against my likely title [share it if you have one].
4. The 3 hook-killers to avoid in my style: long branding intros, 'hey guys welcome back', and throat-clearing context — with my specific temptation named.
5. The transition line from hook into content, so retention does not cliff at second 35.
Act as a thumbnail strategist. My video: [topic and the emotional payoff]. My title draft: [title]. My niche and its thumbnail conventions: [niche].
Lab the thumbnail text:
1. The title-thumbnail split: rule one is they never repeat each other. Show me what the title carries (context, searchability) vs what the thumbnail carries (emotion, curiosity).
2. 10 thumbnail text options, max 4 words each: 3 curiosity, 3 outcome, 2 contrarian, 2 'forbidden fruit'. Each with the facial-expression or visual pairing that completes it.
3. The mobile test: which options survive at 120px wide — cut any that do not.
4. The honesty audit: flag any option that overpromises what my video delivers. Clickbait debt is paid in watch time and trust.
5. Final pick + backup for an A/B swap, with the one-line reasoning a viewer's lizard brain would give for clicking.
Interview questions that get stories, not talking points.
Act as a podcast host known for guests saying 'nobody has ever asked me that'. My guest: [who they are, what they are known for]. My listeners: [who they are, what they hope to take away]. Every interview this guest has already done: [paste links or summarize their standard talking points if known].
Build my bank:
1. The worn path: the 5 questions this guest is always asked, so I can skip or subvert them.
2. 15 original questions across 5 depths: origin story (before the known story starts), craft (the how nobody asks about), failure (specific, not 'tell me about a challenge'), opinion (where they disagree with their own field), and the personal (cost of the journey).
3. For each question: the follow-up that digs one level deeper if they give a polished answer — 'what did that actually look like on the worst day?'
4. The 3 questions to cut if time runs short, keeping the arc.
5. The closer: one unexpected final question that tends to produce the clip everyone shares.