{"slug":"ad-creative","title":"\u5e7f\u544a\u521b\u610f\u751f\u6210","description":"\u751f\u6210\u3001\u8fed\u4ee3\u548c\u89c4\u6a21\u5316\u5e7f\u544a\u521b\u610f\uff1a\u6807\u9898\u3001\u63cf\u8ff0\u3001\u4e3b\u56fe\u6587\u6848\u4e0e\u53d8\u4f53\u3002","category":"\u590d\u76d8\u4e0e\u589e\u957f","tags":["\u5e7f\u544a","\u6295\u653e","\u521b\u610f"],"author":"Corey Haines","source_url":"https://github.com/coreyhaines31/marketingskills/blob/main/skills/ad-creative/SKILL.md","license":"MIT","status":"published","verification":"source_reviewed","verification_notes":"\u6765\u6e90 coreyhaines31/marketingskills\uff08skills/ad-creative/SKILL.md\uff09\uff0c\u672c\u5730\u6253\u5305\u4e0a\u4f20\uff1b\u672a\u6267\u884c\u5305\u5185\u811a\u672c\u3002","revision":"1794c49d32494261a36dba74bc56f288","created_at":"2026-10-28T07:42:54+00:00","updated_at":"2026-10-28T07:44:12+00:00","source_repo":"","source_path":"","source_commit":"","source_updated_at":"","stars":0,"proxy_score":null,"score":null,"grade":null,"dimensions":{},"evaluation":{},"installs":0,"platforms":[],"formats":[],"category_slug":"growth","category_name":"\u590d\u76d8\u4e0e\u589e\u957f","usage_guide":{},"read_url":"https://shiliuai.cn/api/v2/skills/ad-creative","skill_url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/SKILL.md","download_url":"https://shiliuai.cn/api/v2/skills/ad-creative/download","install_url":"https://shiliuai.cn/api/v2/skills/ad-creative/install","files":[{"path":"SKILL.md","bytes":27027,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/SKILL.md"},{"path":"assets/creative-review-template.html","bytes":26092,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/assets/creative-review-template.html"},{"path":"evals/evals.json","bytes":29078,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/evals/evals.json"},{"path":"references/awareness-headlines.md","bytes":6765,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/awareness-headlines.md"},{"path":"references/creative-review-page.md","bytes":9602,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/creative-review-page.md"},{"path":"references/creative-roadmap.md","bytes":9010,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/creative-roadmap.md"},{"path":"references/generative-tools.md","bytes":24215,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/generative-tools.md"},{"path":"references/hook-system.md","bytes":10030,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/hook-system.md"},{"path":"references/imessage-video-ads.md","bytes":19922,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/imessage-video-ads.md"},{"path":"references/meta-creative-formats.md","bytes":10686,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/meta-creative-formats.md"},{"path":"references/motion-video-ads.md","bytes":11801,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/motion-video-ads.md"},{"path":"references/platform-specs.md","bytes":6525,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/platform-specs.md"},{"path":"references/short-form-video-specs.md","bytes":19389,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/short-form-video-specs.md"},{"path":"references/static-ad-templates.md","bytes":23710,"url":"https://shiliuai.cn/api/v2/skills/ad-creative/files/references/static-ad-templates.md"}],"content":"---\nname: ad-creative\ndescription: \"When the user wants to generate, iterate, or scale ad creative \u2014 headlines, descriptions, primary text, or full ad variations \u2014 for any paid advertising platform. Also use when the user mentions 'ad copy variations,' 'ad creative,' 'generate headlines,' 'RSA headlines,' 'bulk ad copy,' 'creative testing,' 'write me some ads,' 'Facebook ad copy,' 'Google ad headlines,' 'LinkedIn ad text,' 'static ads,' 'ad templates,' 'iMessage ad,' 'ChatGPT ad,' 'creative strategy,' 'creative roadmap,' 'creative retro,' 'hook writing,' 'creative review page,' 'motion video ad,' 'faceless video ad,' 'UGC ad,' 'greenscreen ad,' 'TikTok/Reels ad format,' 'Meta ad format tier list,' or 'creative format taxonomy.' Use it to produce or iterate ad copy at scale. Copy avoids AI tells like 'it's not X, it's Y' reveals. For campaign strategy and targeting, see ads. For landing page copy, see copywriting. Test hypotheses must be falsifiable, naming the metric, the comparison, and the expected direction.\"\nmetadata:\n  version: 2.12.2\n---\n\n# Ad Creative\n\nYou are an expert performance creative strategist. Your goal is to generate high-performing ad creative at scale \u2014 headlines, descriptions, and primary text that drive clicks and conversions \u2014 and iterate based on real performance data.\n\n## Before Starting\n\n**Check for product marketing context first:**\nIf `.agents/product-marketing.md` exists (or `.claude/product-marketing.md`, or the legacy `product-marketing-context.md` filename, in older setups), read it before asking questions. Use that context and only ask for information not already covered or specific to this task.\n\nCheck its **Learnings** section before drafting: angles, offers, and formats already tested there shouldn't be re-proposed as new ideas, and active lessons apply unless the user says otherwise.\n\nGather this context (ask if not provided):\n\n### 1. Platform & Format\n- What platform? (Google Ads, Meta, LinkedIn, TikTok, Twitter/X)\n- What ad format? (Search RSAs, display, social feed, stories, video)\n- Are there existing ads to iterate on, or starting from scratch?\n\n### 2. Product & Offer\n- What are you promoting? (Product, feature, free trial, demo, lead magnet)\n- What's the core value proposition?\n- What makes this different from competitors?\n\n### 3. Audience & Intent\n- Who is the target audience?\n- What stage of awareness? (Problem-aware, solution-aware, product-aware)\n- What pain points or desires drive them?\n\n### 4. Performance Data (if iterating)\n- What creative is currently running?\n- Which headlines/descriptions are performing best? (CTR, conversion rate, ROAS)\n- Which are underperforming?\n- What angles or themes have been tested?\n\n### 5. Constraints\n- Brand voice guidelines or words to avoid?\n- Compliance requirements? (Industry regulations, platform policies)\n- Any mandatory elements? (Brand name, trademark symbols, disclaimers)\n\n---\n\n## How This Skill Works\n\nThis skill supports four modes:\n\n### Mode 1: Generate from Scratch\nWhen starting fresh, you generate a full set of ad creative based on product context, audience insights, and platform best practices.\n\n### Mode 2: Iterate from Performance Data\nWhen the user provides performance data (CSV, paste, or API output), you analyze what's working, identify patterns in top performers, and generate new variations that build on winning themes while exploring new angles.\n\nThe core loop:\n\n```\nPull performance data \u2192 Identify winning patterns \u2192 Generate new variations \u2192 Validate specs \u2192 Deliver\n```\n\n### Mode 3: Scaled Static Batches (Grounded)\nFor recurring static ad production at volume (e.g., 50 concepts per batch), work from a **grounded inputs corpus** and the [static ad template library](references/static-ad-templates.md). Every concept must trace to real source material \u2014 see \"Grounded Inputs\" below. To run this on a daily or weekly cadence, see the daily-creative-drop loop in **marketing-loops**. To present a batch for client or stakeholder approval, produce a [creative review page](references/creative-review-page.md).\n\nFor **text-led static headlines**, select treatments by the reader's awareness stage and deliver a mobile-readable visual spec using [references/awareness-headlines.md](references/awareness-headlines.md). It complements the static templates and hook system; numerical claims still need their own evidence.\n\n### Mode 4: Creative Strategy Loop\nFor deciding **which ads are worth making before making them**: synthesize three signal sources (account performance, customer language, external organic) into evidence-ranked concepts, branch the creative mix on account state (exploration vs. scaling), maintain a capacity-checked roadmap with production tiers, and run a monthly retro that feeds the next slate. The full system lives in [references/creative-roadmap.md](references/creative-roadmap.md); for hook generation and funnel-stage diagnosis inside any mode, load [references/hook-system.md](references/hook-system.md).\n\n---\n\n## Grounded Inputs\n\nMost AI ad generation fails on input grounding, not output quality: ungrounded generation produces plausible-sounding ads based on training data, not on what converts for this brand. For scaled production (Mode 3), maintain a durable inputs corpus:\n\n```\ninputs/\n  winning-ads/   10-20 screenshots of the highest-performing ads from the last 90 days\n  reviews/       50-100 customer reviews (Trustpilot, G2, Amazon, App Store) as .md/.txt\n  comments/      Top comments from existing ad campaigns \u2014 objections, unprompted praise, customer-raised angles\nbrand/           Brand voice doc, hex codes, logo, product/screenshot assets\noutputs/         Dated batch folders (outputs/YYYY-MM-DD/)\n```\n\n**Why each input matters:**\n- **Winning ads** carry the hooks, structures, and angles already proven for this brand\n- **Reviews** carry the exact language buyers use for pain, transformation, and unexpected benefits \u2014 pull copy from them verbatim rather than paraphrasing\n- **Ad comments** are the most-skipped and highest-value input: objections (\"but does it work for X?\") become FAQ Card ads, and unprompted praise surfaces angles you didn't write\n\n**Grounding rules:**\n- Every concept cites its source (which review, winning ad, or comment it traces to)\n- No invented claims, stats, or testimonials \u2014 ever\n- If `inputs/winning-ads/` or `inputs/reviews/` is empty, stop and ask the user to populate it before generating. Do not generate ungrounded concepts as a fallback.\n- Inputs decay: refresh `inputs/winning-ads/` as new ads scale; refresh `inputs/reviews/` and `inputs/comments/` monthly\n\n---\n\n## Platform Specs\n\nPlatforms reject or truncate creative that exceeds these limits, so verify every piece of copy fits before delivering.\n\n### Google Ads (Responsive Search Ads)\n\n| Element | Limit | Quantity |\n|---------|-------|----------|\n| Headline | 30 characters | Up to 15 |\n| Description | 90 characters | Up to 4 |\n| Display URL path | 15 characters each | 2 paths |\n\n**RSA rules:**\n- Headlines must make sense independently and in any combination\n- Pin headlines to positions only when necessary (reduces optimization)\n- Include at least one keyword-focused headline\n- Include at least one benefit-focused headline\n- Include at least one CTA headline\n\n### Meta Ads (Facebook/Instagram)\n\n| Element | Limit | Notes |\n|---------|-------|-------|\n| Primary text | 125 chars visible (up to 2,200) | Front-load the hook |\n| Headline | 40 characters recommended | Below the image |\n| Description | 30 characters recommended | Below headline |\n| URL display link | 40 characters | Optional |\n\n### LinkedIn Ads\n\n| Element | Limit | Notes |\n|---------|-------|-------|\n| Intro text | 150 chars recommended (600 max) | Above the image |\n| Headline | 70 chars recommended (200 max) | Below the image |\n| Description | 100 chars recommended (300 max) | Appears in some placements |\n\n### TikTok Ads\n\n| Element | Limit | Notes |\n|---------|-------|-------|\n| Ad text | 80 chars recommended (100 max) | Above the video |\n| Display name | 40 characters | Brand name |\n\n### Twitter/X Ads\n\n| Element | Limit | Notes |\n|---------|-------|-------|\n| Tweet text | 280 characters | The ad copy |\n| Headline | 70 characters | Card headline |\n| Description | 200 characters | Card description |\n\nFor detailed specs and format variations, see [references/platform-specs.md](references/platform-specs.md).\n\n---\n\n## Generating Ad Visuals\n\n**To decide *which format to make next*** (before briefing any specific ad), consult the Meta creative format taxonomy in [references/meta-creative-formats.md](references/meta-creative-formats.md) \u2014 a prioritized S\u2192F catalog of ~51 formats ranked by one question: is it a *unicorn scaler* that punctures cold net-new audiences, or a *supporting cast* member that only converts mid-funnel? Leads with the persona-based Andromeda context (why creator-fronted formats top the list), S-tier callouts (founder content, partnership ads, VSL), the A-tier bench, and explicit F-tier de-prioritization (press, podcast, notes-app fake-native). Use it to pick a format and build a portfolio; the how-to-build detail lives in the static/video references below. For the account-level kill/keep/scale math once ads are live, cross-reference the `ads` skill's the **ads** skill's Meta decision system reference.\n\n**For static ad structure**, use the template library in [references/static-ad-templates.md](references/static-ad-templates.md) \u2014 layout frameworks (Us vs. Them, Stat Callout, Review Card, Before/After, Founder Message, FAQ Card, Grid Static, Callout, and more) with copy slots, DTC and SaaS examples, and per-concept output format. Each template carries a **tier (S\u2013F)** and **funnel role** (unicorn cold-scaler vs. mid-funnel supporting cast) so you reach for the right one first. Cycle through templates rather than clustering on favorites \u2014 but weight toward the S/A tiers when the goal is cold net-new reach.\n\n**For iOS-native reveal video ads** \u2014 iMessage chat reveals (scripted thread unfolds bubble-by-bubble: screenshot hook \u2192 friend asks \"what app is that?\" \u2192 brand + promo code reveal \u2192 end card), ChatGPT reveals (typed question \u2192 streaming answer), Apple Notes reveals (a confessional note typed live), and AirDrop reveals (an incoming share where the accept-tap is the reveal) \u2014 see [references/imessage-video-ads.md](references/imessage-video-ads.md) for surface selection, the six concept angles, script and pacing rules, production routes (Playwright + ffmpeg pipeline, code-based video), craft details that sell the illusion, and the grounding/compliance rules for dramatized conversations (strictest for fabricated AI answers).\n\n**For faceless motion-style video ads** \u2014 fully generated 15\u201345s concept/explainer videos (styled poster stills \u2192 image-to-video \"living\" motion \u2192 TTS narration \u2192 word-timed captions; roughly $3\u20136 and ~15 minutes per finished video) \u2014 see [references/motion-video-ads.md](references/motion-video-ads.md) for the provider-agnostic pipeline, a nine-style visual library with fill-in prompt formulas \u2014 five characterful looks (screen-print collage, flat vector explainer, papercraft diorama, pop-art comic, claymation) plus four brand-flexible token-driven styles (monoline editorial, Swiss typographic, wireglow, duotone screenprint) driven by a brand-slots contract (FIELD / INK / ACCENT / TYPE FEEL) \u2014 the motion prompt formula, and hard-earned QC gotchas (maker-hands intrusion, final-two-seconds drift, caption/label collision, TTS/whisper sound-alikes).\n\n**For creator/UGC short-form video** \u2014 a tiered format library (reaction+demo hard cuts, \"no yapping\" split-screen tutorials, greenscreen reactions, plus Yapper, amateur investigation, David & Goliath, authority, VSL, green-screen commentary, conversation, duet/reaction, ASMR, and street-interview formats, each with a scale-vs-support tier and mechanics) and founder / organic-vlog structures (hero's journey, math, shiny-object, niche-guide, the three-capture shooting system, and the 0.5\u20131s cut formula) for TikTok/Reels/Shorts growth and paid \u2014 see [references/short-form-video-specs.md](references/short-form-video-specs.md). It also carries the **vertical video production spec** that applies to *all* 9:16 video this skill makes: the cross-platform safe-zone band (720\u00d71200 text-safe area \u2014 the most-missed constraint), the classic TikTok caption recipe (white fill + black stroke, no pill), static-caption auto-sizing, and the organic-vs-baked-music decision that affects reach. Load it before producing any vertical video.\n\nFor image and video generation tools, see [references/generative-tools.md](references/generative-tools.md) for the complete guide covering:\n\n- **Image generation** \u2014 Nano Banana Pro (Gemini), Flux, Ideogram for static ad images\n- **Video generation** \u2014 Veo, Kling, Runway, Seedance, Higgsfield for video ads\n- **Voice & audio** \u2014 ElevenLabs, OpenAI TTS, Cartesia for voiceovers, cloning, multilingual\n- **Code-based video** \u2014 Remotion or Hyperframes for templated, data-driven video at scale\n- **Platform image specs** \u2014 Correct dimensions for every ad placement\n- **Cost comparison** \u2014 Pricing for 100+ ad variations across tools\n\n**Workflow for scaled production:**\n1. Generate hero creative with AI tools (exploratory, high-quality)\n2. Build code-based video templates (Remotion or Hyperframes) based on winning patterns\n3. Batch produce variations from data feeds\n4. Iterate \u2014 AI for new angles, templates for scale\n\n---\n\n## Generating Ad Copy\n\n### Step 1: Define Your Angles\n\nBefore writing individual headlines, establish 3-5 distinct **angles** \u2014 different reasons someone would click. Each angle should tap into a different motivation.\n\n**Common angle categories:**\n\n| Category | Example Angle |\n|----------|---------------|\n| Pain point | \"Stop wasting time on X\" |\n| Outcome | \"Achieve Y in Z days\" |\n| Social proof | \"Join 10,000+ teams who...\" |\n| Curiosity | \"The X secret top companies use\" |\n| Comparison | \"Unlike X, we do Y\" |\n| Urgency | \"Limited time: get X free\" |\n| Identity | \"Built for [specific role/type]\" |\n| Contrarian | \"Why [common practice] doesn't work\" |\n\n### Step 2: Generate Variations per Angle\n\nFor each angle, generate multiple variations. Vary:\n- **Word choice** \u2014 synonyms, active vs. passive\n- **Specificity** \u2014 numbers vs. general claims\n- **Tone** \u2014 direct vs. question vs. command\n- **Structure** \u2014 short punch vs. full benefit statement\n\n### Step 3: Validate Against Specs\n\nBefore delivering, check every piece of creative against the platform's character limits. Flag anything that's over and provide a trimmed alternative.\n\n### Step 4: Organize for Upload\n\nPresent creative in a structured format that maps to the ad platform's upload requirements.\n\n---\n\n## Iterating from Performance Data\n\nWhen the user provides performance data, follow this process:\n\n### Step 1: Analyze Winners\n\nLook at the top-performing creative (by CTR, conversion rate, or ROAS \u2014 ask which metric matters most) and identify:\n\n- **Winning themes** \u2014 What topics or pain points appear in top performers?\n- **Winning structures** \u2014 Questions? Statements? Commands? Numbers?\n- **Winning word patterns** \u2014 Specific words or phrases that recur?\n- **Character utilization** \u2014 Are top performers shorter or longer?\n\n### Step 2: Analyze Losers\n\nLook at the worst performers and identify:\n\n- **Themes that fall flat** \u2014 What angles aren't resonating?\n- **Common patterns in low performers** \u2014 Too generic? Too long? Wrong tone?\n\n### Step 3: Generate New Variations\n\nCreate new creative that:\n- **Doubles down** on winning themes with fresh phrasing\n- **Extends** winning angles into new variations\n- **Tests** 1-2 new angles not yet explored\n- **Avoids** patterns found in underperformers\n\n### Step 4: Document the Iteration\n\nTrack what was learned and what's being tested:\n\n```\n## Iteration Log\n- Round: [number]\n- Date: [date]\n- Top performers: [list with metrics]\n- Winning patterns: [summary]\n- New variations: [count] headlines, [count] descriptions\n- New angles being tested: [list]\n- Angles retired: [list]\n```\n\n---\n\n## Writing Quality Standards\n\n### Headlines That Click\n\n**Strong headlines:**\n- Specific (\"Cut reporting time 75%\") over vague (\"Save time\")\n- Benefits (\"Ship code faster\") over features (\"CI/CD pipeline\")\n- Active voice (\"Automate your reports\") over passive (\"Reports are automated\")\n- Include numbers when possible (\"3x faster,\" \"in 5 minutes,\" \"10,000+ teams\")\n\n**Avoid:**\n- Jargon the audience won't recognize\n- Claims without specificity (\"Best,\" \"Leading,\" \"Top\")\n- All caps or excessive punctuation\n- Clickbait that the landing page can't deliver on\n- Shortened or nicknamed product names \u2014 use the exact product name the\n  brief gives, in every field of every concept, unless a short form is\n  explicitly approved\n\n### Test Hypotheses That Can Lose\n\nWhen a concept ships with a test hypothesis, write one that a result can\nprove wrong. \"This will resonate with renters\" cannot lose; \"beats the\ncontrol on CTR among renters\" can.\n\n- Name the metric (CTR, CVR, CPA, thumb-stop rate) and the comparison \u2014\n  against which control or sibling concept, in which audience\n- State the expected direction, and a magnitude when a baseline exists\n  (\"+15% CTR vs. Concept A\")\n- The hypothesis should follow from the angle's motivation, so a win or a\n  loss teaches you something about the audience, not just the asset\n\n### No AI Tells\n\nAd copy that reads as generated looks like every other ad in the feed, and generating at volume multiplies whatever tic the first draft has. Check the whole batch.\n\nNever write these:\n- **Contrast reveals**: \"It's not X, it's Y.\" \"Not because X. Because Y.\" State Y directly, with the reason.\n- **Negation lists**: \"No X, no Y, no Z.\" Say what does happen. One plain absence (\"No card required\") is fine.\n- **Trailing pile-ons**: a full claim, then a comma and more restating clauses. End the sentence at the claim.\n- **Self-answered questions and colon reveals**: \"The result? 3x faster.\" \"The best part: it learns.\" FAQ questions and a question in the reader's own voice are fine.\n- **Stock phrases**: \"Say goodbye to,\" \"X, reimagined,\" \"Unlock the power of,\" \"Take it to the next level,\" \"Here's the thing.\"\n- **Em dashes** in headlines, descriptions, and primary text.\n\nAd-specific tells:\n- **Stacked negations to fit a character limit**: \"No setup. No fees. No contracts.\" Keep the one that matters most and use the space for a fact.\n- **Hook slop**: a generic \"POV:\", \"Stop scrolling,\" \"Here's what nobody tells you about X.\" Earn the next second with something specific to the viewer. A POV hook is fine when the situation is hyper-specific (\"POV: it's 3pm and you're on your fourth coffee\").\n- **CTA slop**: \"Get started today,\" \"Join thousands of happy customers.\" Name the action and what they get.\n\nAcross a batch, vary the sentence shapes. Ten variations that all open the same way read as generated, even when each one is fine alone. For the full blacklist, use the **copywriting** skill's AI-tells reference.\n\n### Descriptions That Convert\n\nDescriptions should complement headlines, not repeat them. Use descriptions to:\n- Add proof points (numbers, testimonials, awards)\n- Handle objections (\"No credit card required,\" \"Free forever for small teams\")\n- Reinforce CTAs (\"Start your free trial today\")\n- Add urgency when genuine (\"Limited to first 500 signups\")\n\n---\n\n## Output Formats\n\n### Standard Output\n\nOrganize by angle, with character counts:\n\n```\n## Angle: [Pain Point \u2014 Manual Reporting]\n\n### Headlines (30 char max)\n1. \"Stop Building Reports by Hand\" (29)\n2. \"Automate Your Weekly Reports\" (28)\n3. \"Reports Done in 5 Min, Not 5 Hr\" (31) <- OVER LIMIT, trimmed below\n   -> \"Reports in 5 Min, Not 5 Hrs\" (27)\n\n### Descriptions (90 char max)\n1. \"Marketing teams save 10+ hours/week with automated reporting. Start free.\" (73)\n2. \"Connect your data sources once. Get automated reports forever. No code required.\" (80)\n```\n\n### Bulk CSV Output\n\nWhen generating at scale (10+ variations), offer CSV format for direct upload:\n\n```csv\nheadline_1,headline_2,headline_3,description_1,description_2,platform\n\"Stop Manual Reporting\",\"Automate in 5 Minutes\",\"Join 10K+ Teams\",\"Save 10+ hrs/week on reports. Start free.\",\"Connect data sources once. Reports forever.\",\"google_ads\"\n```\n\n### Static Batch Output (Mode 3)\n\nFor scaled static batches, save to a dated folder with an index:\n\n```\noutputs/YYYY-MM-DD/\n  INDEX.md        # every concept: template type + grounding source, scannable in 2 min\n  concepts/       # one .md per concept: headline, body, visual description, image prompt, grounding\n  images/         # generated images, if an image tool is configured\n```\n\nPer-concept format is defined in [references/static-ad-templates.md](references/static-ad-templates.md). The human workflow this supports: open the folder, scan INDEX.md, pick the best 5-10 for testing \u2014 picking 5 winners from 50 concepts yields better creative than picking 5 from 10.\n\n### Creative Review Page (client / stakeholder approval)\n\nWhen a person who isn't you needs to review and pick \u2014 a client, a partner, a stakeholder \u2014 produce a **creative review page**: a self-contained HTML artifact that presents each concept as an in-feed platform mockup (Instagram/Facebook, with a whitelist-handle toggle), breaks carousels into a labeled frame-by-frame storyboard, lets them toggle headline/copy variations, and discloses what's grounded in real assets. It's the visual upgrade to INDEX.md \u2014 a decision made off one link instead of by reading markdown. The template ships at [assets/creative-review-template.html](assets/creative-review-template.html) (one file, no build, hostable anywhere); populate its `DATA` object from your generated concepts. Full data model, grounding rules (the disclosure block is required), and delivery in [references/creative-review-page.md](references/creative-review-page.md).\n\n### Iteration Report\n\nWhen iterating, include a summary:\n\n```\n## Performance Summary\n- Analyzed: [X] headlines, [Y] descriptions\n- Top performer: \"[headline]\" \u2014 [metric]: [value]\n- Worst performer: \"[headline]\" \u2014 [metric]: [value]\n- Pattern: [observation]\n\n## New Creative\n[organized variations]\n\n## Recommendations\n- [What to pause, what to scale, what to test next]\n```\n\n---\n\n## Batch Generation Workflow\n\nFor large-scale creative production (Anthropic's growth team generates 100+ variations per cycle):\n\n### 1. Break into sub-tasks\n- **Headline generation** \u2014 Focused on click-through\n- **Description generation** \u2014 Focused on conversion\n- **Primary text generation** \u2014 Focused on engagement (Meta/LinkedIn)\n\n### 2. Generate in waves\n- Wave 1: Core angles (3-5 angles, 5 variations each)\n- Wave 2: Extended variations on top 2 angles\n- Wave 3: Wild card angles (contrarian, emotional, specific)\n\n### 3. Quality filter\n- Remove anything over character limit\n- Remove duplicates or near-duplicates\n- Flag anything that might violate platform policies\n- Ensure headline/description combinations make sense together\n\n---\n\n## Common Mistakes\n\n- **Writing headlines that only work together** \u2014 RSA headlines get combined randomly\n- **Ignoring character limits** \u2014 Platforms truncate without warning\n- **All variations sound the same** \u2014 Vary angles, not just word choice\n- **No CTA headlines** \u2014 RSAs need action-oriented headlines to drive clicks; include at least 2-3\n- **Generic descriptions** \u2014 \"Learn more about our solution\" wastes the slot\n- **Iterating without data** \u2014 Gut feelings are less reliable than metrics\n- **Generating without grounding** \u2014 Ungrounded concepts read like every other ad in the feed; feed the skill winning ads, reviews, and comments first\n- **Skipping the comments input** \u2014 Ad comments hold the objections and angles customers raise themselves; those usually convert best\n- **Testing too many things at once** \u2014 Change one variable per test cycle\n- **Retiring creative too early** \u2014 Allow 1,000+ impressions before judging\n\n---\n\n## Tool Integrations\n\nFor pulling performance data and managing campaigns, see the [tools registry](https://github.com/coreyhaines31/marketingskills/blob/main/tools/REGISTRY.md).\n\n| Platform | Pull Performance Data | Manage Campaigns | Guide |\n|----------|:---------------------:|:----------------:|-------|\n| **Google Ads** | `google-ads campaigns list`, `google-ads reports get` | `google-ads campaigns create` | [google-ads.md](https://github.com/coreyhaines31/marketingskills/blob/main/tools/integrations/google-ads.md) |\n| **Meta Ads** | `meta-ads insights get` | `meta-ads campaigns list` | [meta-ads.md](https://github.com/coreyhaines31/marketingskills/blob/main/tools/integrations/meta-ads.md) |\n| **LinkedIn Ads** | `linkedin-ads analytics get` | `linkedin-ads campaigns list` | [linkedin-ads.md](https://github.com/coreyhaines31/marketingskills/blob/main/tools/integrations/linkedin-ads.md) |\n| **TikTok Ads** | `tiktok-ads reports get` | `tiktok-ads campaigns list` | [tiktok-ads.md](https://github.com/coreyhaines31/marketingskills/blob/main/tools/integrations/tiktok-ads.md) |\n\n### Workflow: Pull Data, Analyze, Generate\n\n```bash\n# 1. Pull recent ad performance\nnode tools/clis/google-ads.js reports get --type ad_performance --date-range last_30_days\n\n# 2. Analyze output (identify top/bottom performers)\n# 3. Feed winning patterns into this skill\n# 4. Generate new variations\n# 5. Upload to platform\n```\n\n> **Skip the plumbing.** Step 1 is the hard part to run yourself \u2014 the Google\n> Ads (or Meta / LinkedIn / TikTok) API, OAuth, and a reporting query per\n> account, repeated every time. [Magister](https://magistermarketing.com?ref=marketingskills)\n> is an autonomous marketing agent that runs on these exact skills and pulls\n> ad performance natively: its agent reads your accounts through typed,\n> approval-gated integrations with no setup, then applies this skill to the\n> result. If you're wiring these pulls up by hand, it's worth a look.>\n> *Disclosure: Magister is built by Corey Haines, who maintains this repository.*\n> To run the pull yourself instead, use the repo's ad platform CLIs (setup in\n> [tools/SETUP.md](https://github.com/coreyhaines31/marketingskills/blob/main/tools/SETUP.md)),\n> or a reporting connector such as Supermetrics or Coupler.\n\n---\n\n## Related Skills\n\n- **ads**: For campaign strategy, targeting, budgets, and optimization\n- **marketing-loops**: For running static batch generation on a recurring cadence (the daily-creative-drop loop)\n- **customer-research**: For mining reviews and comments when building the grounded inputs corpus\n- **copywriting**: For landing page copy (where ad traffic lands)\n- **ab-testing**: For structuring creative tests with statistical rigor\n- **marketing-psychology**: For psychological principles behind high-performing creative\n- **copy-editing**: For polishing ad copy before launch\n","usage":"\u8bfb\u53d6 SKILL.md\uff0c\u5e76\u6309\u9700\u8bfb\u53d6\u914d\u5957\u6587\u4ef6\uff1b\u4e0b\u8f7d ZIP \u83b7\u53d6\u5b8c\u6574\u5305\u3002\u6280\u80fd\u662f\u4efb\u52a1\u65b9\u6cd5\uff0c\u4e0d\u6388\u4e88\u65b0\u589e\u8d26\u6237\u3001\u4ed8\u8d39\u6216\u7cfb\u7edf\u6743\u9650\u3002"}