ads-hooks

# Scroll-Stopping Hook Generator > **Portability note:** This skill writes its deliverable to `ADS-HOOKS.md` in the current working directory and references sibling `/ads *` commands. On non-OmegaOS hosts, treat `/ads copy`, `/ads video`, and `/ads strategy` as optional follow-ups — the hook generation itself is fully self-contained and needs only `WebFetch`/`WebSearch` (skip gracefully if unavailable and ask the user to paste business context). ## Single-voice craft (do NOT parallelize) Hooks are a **single coherent persuasive voice**, not a multi-angle audit. The five psychological angles are facets of ONE brand voice talking to ONE audience — they must sound like the same sharp copywriter, not five disconnected agents. **Do NOT fan this out across parallel sub-agents or workflows.** Splitting the work fragments voice, repeats phrasing, and destroys the rhythm that makes a hook stop the scroll. Generate all hooks in one pass, in one voice, then self-edit as a single craftsman. **Craft upgrades (apply to every hook):** 1. **Voice lock before writing.** In one line, define the voice (e.g. "blunt founder", "warm expert", "rebel insider") from the business context, then keep EVERY hook in that register. Mixed registers = amateur tell. 2. **Specificity over adjectives.** Replace every vague claim with a concrete number, timeframe, or named detail ("Save $347/month", not "save money"). If a hook has no specific, it is not done. 3. **Read-aloud rhythm pass.** Say each hook out loud. Cut filler words, front-load the surprise, end on the highest-tension word. Awkward to say = rewrite. The best hooks are conversational, not literary. ## Skill Purpose Generate 20+ battle-tested hooks that stop the scroll across every major ad platform. Hooks are the single most important element in any ad — the first 1-3 seconds determine whether your ad gets watched or ignored. This skill produces hooks organized by psychological angle, each with platform-specific formatting, an A/B variation, a psychology breakdown explaining why it works, and guidance on where to use it. Output is ready to plug directly into ad copy, video scripts, or email subject lines. ## When to Use - User runs `/ads hooks` - User asks for scroll-stopping hooks, opening lines, or attention-grabbers - Called as a subagent from `/ads strategy` (the main orchestrator) - User needs hooks for ad creative, video intros, or email campaigns - User wants to improve CTR on existing ads by testing new hooks ## Input Requirements - **Required:** Business context — a URL, product/service description, or prior strategy data - **Optional:** Target persona, specific platform focus, industry vertical, offer details, reference or competitor video URLs (feed them to /watch first) ## How to Execute ### Step 0 (optional): Watch the reference videos first If the user can point to reference or competitor videos (YouTube, TikTok, Instagram, Loom, X links, or local files), run /watch on 2 or 3 of them before writing anything. The hook microscope in its report.md dissects the 0-10s opening frame by frame: the exact first words, the on-screen text, the first visual change. Use those observations as the hook brief: extract the patterns that make the winning openings work (pain called out in second one? bold claim? pattern interrupt?), then write hooks grounded in observed behavior instead of generic formulas. Skip this step when no reference videos exist. ### Step 1: Gather Business & Audience Context If a URL is provided, fetch it using `WebFetch` and extract: - Business name, industry, and niche - Core value proposition and USPs - Target audience indicators - Pain points the product/service solves - Price point and offer structure - Social proof assets (review count, client count, awards) - Competitor landscape Run supplementary research: ``` WebSearch: "[Industry]" best performing ad hooks WebSearch: "[Industry]" viral ad examples WebSearch: "[Product category]" ad copy examples high CTR ``` Establish the hook foundation: ```markdown ## Hook Foundation **Business:** [Name] **Product/Service:** [What they sell] **Target Audience:** [Primary persona in 1 sentence] **Core Pain Point:** [The #1 problem the audience has] **Core Desire:** [The #1 outcome the audience wants] **USP:** [What makes this different from competitors] **Proof Points:** [Stats, results, testimonials available] **Offer:** [Current promotion, pricing, free trial] ``` ### Step 2: Generate Hooks by Psychological Angle Generate **20+ hooks minimum** distributed across 5 psychological angles: --- #### Angle 1: Pain/Problem Hooks (5 minimum) These hooks work by calling out a pain the audience already feels. They create instant recognition — "that's me!" — which stops the scroll. **Psychology:** Loss aversion. People are 2x more motivated to avoid pain than to gain pleasure. Pain hooks activate the amygdala's threat detection system. **Template patterns:** - "Stop [doing painful thing] and start [doing desired thing]" - "If you're still [painful situation], here's why..." - "[Number]% of [audience] are making this [topic] mistake" - "The [topic] mistake that's costing you [specific loss]" - "You're losing [specific amount] every [time period] because..." For each hook: ```markdown ### Hook P[number]: [Hook Text] **Full text:** [The complete hook — 1-2 sentences max] **A/B variation:** [Alternate version testing a different angle or word choice] **Character count:** [count] characters **Best platform(s):** [where this hook performs best] **Funnel stage:** [TOFU / MOFU / BOFU] **Why it works:** [2-3 sentences on the psychology — which cognitive bias, emotional trigger, or pattern interrupt is at play] **How to use it:** - **Facebook/Instagram ad:** [how to format — as primary text opening, headline, or text overlay] - **TikTok/Reels:** [how to deliver — on-screen text, voiceover opening, or both] - **Google ad:** [how to adapt — as headline, description opening] - **YouTube pre-roll:** [how to deliver in first 5 seconds] - **Email subject line:** [adapted version for email] **Emoji treatment:** [which emojis to add and where, or "none — let the words hit"] **Hashtag pairing:** [2-3 relevant hashtags if used on social] ``` --- #### Angle 2: Curiosity/Question Hooks (5 minimum) These hooks create an information gap that the viewer must close. The brain cannot tolerate an open loop — curiosity is physically uncomfortable until resolved. **Psychology:** The Zeigarnik Effect. Unfinished thoughts create cognitive tension. Question hooks exploit the brain's compulsion to find answers. **Template patterns:** - "What happens when you [unexpected action] with [topic]?" - "I tested [thing] for [time period]. Here's what happened..." - "The [topic] secret that [authority] don't want you to know" - "Why do [successful people] always [unexpected behavior]?" - "I asked [number] [experts] one question. Their answer shocked me." - "This [simple thing] changed my [topic] forever. (Not what you think.)" For each hook, use the same template as Angle 1. --- #### Angle 3: Social Proof/Authority Hooks (5 minimum) These hooks leverage other people's actions or endorsements to trigger herd behavior. If others are doing it, it must be good. **Psychology:** Social proof (Cialdini). Authority bias. The bandwagon effect. We look to others' behavior to determine the correct course of action, especially under uncertainty. **Template patterns:** - "[Number] [audience type] already [action]. Here's why." - "[Authority figure/company] uses this [product/strategy]. You should too." - "We helped [X] clients [achieve result] in [timeframe]" - "Rated [X] stars by [number] [audience]. Here's what they're saying..." - "[Well-known company] switched to [product/approach]. Their results?" - "The [product/method] that [impressive stat] people can't stop talking about" For each hook, use the same template as Angle 1. --- #### Angle 4: Contrarian/Controversial Hooks (3 minimum) These hooks challenge conventional wisdom or say something the audience doesn't expect. They create a pattern interrupt — the brain flags unexpected information as important. **Psychology:** The Von Restorff effect (isolation effect). Information that stands out from its surroundings is remembered better. Contrarian takes also trigger the brain's prediction error response, releasing dopamine and increasing attention. **Template patterns:** - "[Common advice] is actually killing your [metric]. Here's proof." - "Unpopular opinion: [contrarian take about industry]" - "Everything you've been told about [topic] is wrong." - "I stopped [common practice] and [positive result] happened." - "The [industry] lie nobody talks about" For each hook, use the same template as Angle 1. **Important:** Contrarian hooks must be defensible. Never make false claims. The contrarian take should challenge a genuine misconception or offer a legitimate alternative perspective. --- #### Angle 5: Urgency/Scarcity Hooks (2 minimum) These hooks create time pressure or exclusivity. They work best at the bottom of the funnel when the audience is already aware and considering. **Psychology:** Scarcity principle (Cialdini) and loss aversion (Kahneman). The fear of missing out (FOMO) is a stronger motivator than the joy of gaining. Time-bound offers trigger the brain's urgency response, shortcutting deliberation. **Template patterns:** - "Last chance: [offer] ends [specific time]" - "Only [number] spots left for [offer/program]" - "[Offer] expires in [time]. Don't say I didn't warn you." - "This [deal/opportunity] won't come back. Here's why." - "We're closing [offer] to new [customers/members] on [date]" For each hook, use the same template as Angle 1. **Important:** Urgency hooks must be truthful. False scarcity destroys trust and violates platform policies. Only use when there is genuine scarcity or a real deadline. ### Step 3: Platform-Specific Formatting Guide Include a comprehensive formatting reference: ```markdown ## Platform-Specific Hook Formatting ### Facebook/Instagram Feed Ads - **Character guidance:** Keep hooks under 125 characters to display above the fold (before "See More") - **Emoji usage:** 1-2 emojis max at the start or end — never in the middle of a sentence - **Line breaks:** Use a line break after the hook to create visual separation - **Hashtags:** 0 in ad copy (use in organic posts only, 3-5 max) - **Format:** Text-first, then visual reinforcement ### TikTok / Instagram Reels - **Delivery:** Hook must be spoken in the first 1.5 seconds AND shown as on-screen text - **Text overlay:** Large, bold, centered text — 5-8 words max on screen - **Emoji usage:** Sparingly in on-screen text, more in captions - **Hashtags:** 3-5 in caption, include 1 trending hashtag - **Format:** Video-first, text supports the visual ### Google Ads (Search) - **Headline limit:** 30 characters per headline — hooks must be compressed - **No emojis:** Google Ads does not allow emojis in ad copy - **Keyword integration:** Include the target keyword naturally in the hook - **Punctuation:** Use periods and exclamation points strategically (Google shows them) - **Format:** Intent-matching — the hook must answer what they searched for ### LinkedIn - **Tone:** Professional but human — avoid jargon, corporate-speak, and buzzwords - **Emoji usage:** Minimal (0-1 per post) — overuse damages credibility - **Hashtags:** 3-5 in Sponsored Content, placed at the end - **Format:** Text hooks work best, followed by a supporting statistic - **Line breaks:** Use frequently — LinkedIn's feed compresses long paragraphs ### YouTube Pre-Roll - **Timing:** Hook must land in the first 5 seconds (before the skip button) - **Delivery:** Both spoken (voiceover) and shown (on-screen text) - **No emojis:** On-screen text should be clean and bold - **Hashtags:** Not applicable - **Format:** Open with the hook verbally, reinforce with on-screen text and a relevant visual ### Pinterest - **Placement:** Hook goes in the Pin title (100 chars) and the first line of description - **Emoji usage:** Moderate — Pinterest is more visual, emojis help text stand out - **Hashtags:** 2-5 in description, relevant to search intent - **Format:** Keyword-rich — Pinterest is a search engine, hooks should include searchable terms ### Email Subject Lines - **Character guidance:** 40-60 characters for desktop, under 35 for mobile preview - **Emoji usage:** 1 emoji at the start can increase open rates by 10-15% - **Personalization:** Include [First Name] or [Company] where possible - **Preview text:** 40-90 characters that complement (not repeat) the subject line - **Format:** Create curiosity gap — never give away the answer in the subject line ``` ### Step 4: Hook Performance Prediction Rate each hook's expected performance: ```markdown ## Hook Performance Predictions | # | Hook (shortened) | Angle | Expected CTR Impact | Confidence | Best Platform | |---|---|---|---|---|---| | P1 | [first 40 chars...] | Pain | High (+30-50%) | Medium | Meta | | P2 | [first 40 chars...] | Pain | Medium (+15-30%) | High | Google | | C1 | [first 40 chars...] | Curiosity | Very High (+50%+) | Medium | TikTok | | [etc.] | | | | | | **CTR Impact Rating:** - Very High: Expected to outperform average hooks by 50%+ - High: Expected to outperform by 30-50% - Medium: Expected to outperform by 15-30% - Baseline: Expected to perform at industry average **Confidence Rating:** - High: Based on proven patterns and strong audience match - Medium: Based on solid principles but needs testing - Low: Experimental angle, high variance expected ``` ### Step 5: Hook Testing Strategy ```markdown ## A/B Testing Strategy ### Phase 1: Angle Testing (Week 1-2) Test one hook from each angle against each other: - Ad Set A: [Best Pain hook] - Ad Set B: [Best Curiosity hook] - Ad Set C: [Best Social Proof hook] - Ad Set D: [Best Contrarian hook] - Budget: Equal split across all 4 ad sets - Metric: CTR (click-through rate) - Winner criteria: Highest CTR after 1,000 impressions per ad set ### Phase 2: Variation Testing (Week 3-4) Take the winning angle and test A/B variations: - Ad A: [Winning hook — Version A] - Ad B: [Winning hook — Version B] - Budget: 50/50 split - Metric: CTR + conversion rate - Winner criteria: Highest conversion rate after 500 clicks per ad ### Phase 3: Scaling (Week 5+) Take the overall winner and: - Scale budget by 20% every 3 days - Create 3 new hooks in the winning angle - Refresh creative every 2-3 weeks to avoid ad fatigue - Monitor frequency — if above 3.0, rotate new hooks in ``` ## Output Format Save the complete output to `ADS-HOOKS.md` in the current working directory. **File structure:** ``` # Scroll-Stopping Hooks: [Business Name] > Generated [date] | Source: [URL] > Total hooks: [count] | Angles covered: 5 > Top recommended hook: [Hook #] — [shortened text] ## Hook Foundation [business and audience context] ## Pain/Problem Hooks (5) [hooks P1 through P5 with full detail] ## Curiosity/Question Hooks (5) [hooks C1 through C5 with full detail] ## Social Proof/Authority Hooks (5) [hooks S1 through S5 with full detail] ## Contrarian/Controversial Hooks (3) [hooks X1 through X3 with full detail] ## Urgency/Scarcity Hooks (2) [hooks U1 through U2 with full detail] ## Platform-Specific Formatting Guide [formatting reference] ## Hook Performance Predictions [prediction table] ## A/B Testing Strategy [testing plan] ## Quick-Reference Hook List [numbered list of all hooks — just the text, no formatting, for easy scanning] ## Next Steps 1. Start with hooks [X, Y, Z] — highest predicted performance 2. Run `/ads copy <platform>` to build full ad copy around winning hooks 3. Run `/ads video <product>` to turn hooks into video ad scripts 4. Set up A/B tests following the Phase 1 plan above 5. Refresh hooks every 3-4 weeks to prevent creative fatigue ``` ## Output Contract & Verify **Produces:** exactly one file — `ADS-HOOKS.md` in the current working directory — containing ≥20 hooks across all 5 angles, each with A/B variation, character count, platform usage, and psychology breakdown, plus the formatting guide, performance-prediction table, A/B testing plan, and quick-reference list. **Verify step (run before claiming done):** confirm the file exists and `grep -c "^### Hook"` (or a manual count) returns ≥20. Walk the Quality Checklist below item-by-item against the actual file content — not from memory. If any item fails, fix it before delivering. **Evidence / no-hallucination guardrail:**