1. ./brands/{brand-slug}/brand-context.md

   -- brand identity, audience, USP

2. ./brands/{brand-slug}/product-marketing-context.md

   -- deep positioning, customer language, objections (read if it exists)

3. ./brands/{brand-slug}/sostac/01-situation.md

   -- current state, budget, competitors

4. ./brands/{brand-slug}/sostac/02-objectives.md

   -- goals, KPIs, targets

5. ./brands/{brand-slug}/sostac/03-strategy.md

   -- target segments, positioning, phasing

6. ./brands/{brand-slug}/sostac/04-tactics.md

   -- channel plan, paid ads role, budget allocation

./brands/{brand-slug}/sostac/00-auto-discovery.md

agent-browser

Setup: Before running research, check if

agent-browser

is available (

agent-browser --version

). If the command is not found, install it:

npm install -g agent-browser && npx playwright install chromium

. If installation fails, use

WebFetch

and

WebSearch

tools as alternatives for all research tasks in this section.

• Keyword Strategy: Build from SOSTAC audience language. Group keywords into tight ad groups by theme and intent. Use phrase match as default, exact match for high-value terms, broad match only with smart bidding and strong conversion data.
• Negative Keywords: Build proactive negative lists (competitors, irrelevant modifiers, free/cheap if premium brand, job-related terms). Review search terms weekly. Maintain account-level and campaign-level negative lists.
• Ad Copy: 15 headlines (mix keyword, benefit, CTA, social proof, urgency), 4 descriptions. Pin sparingly. Include primary keyword in H1 and H2.
• Ad Extensions: Sitelinks (4-6 with descriptions), callouts (unique selling points), structured snippets (product types, services), call, location, price, promotion extensions. Use all that apply.
• Quality Score: Target 7+ on core keywords. Improve via tighter keyword-ad-landing page relevance, higher expected CTR (better copy), and landing page experience (speed, mobile, relevance).

• Asset Groups: Organize by product line or audience segment, not lumped together. Provide 20+ images, 5+ videos, 15 headlines, 5 long headlines, 5 descriptions, business name, final URL.
• Audience Signals: Layer custom segments (search terms, competitor URLs), your data (customer lists, converters), demographics, and interests. Signals guide the algorithm, they do not restrict it.
• Feed Optimization: Title = brand + product + key attribute + variant. Clean descriptions. GTINs. High-quality images. Supplemental feeds for custom labels and promotions.
• Monitoring: Break down by asset group performance. Watch for cannibalization of brand search. Use scripts to pull placement reports. Exclude low-quality placements.

• Display: Managed placements for brand safety. Responsive display ads with 5+ images and 5+ headlines. Layer audience targeting (in-market, custom intent, remarketing) over topic/placement targeting.
• Demand Gen: Replaces Discovery. Runs across YouTube, Discover, Gmail. Provide video + image assets. Use lookalike segments based on converters.

• In-Stream (Skippable): Hook in first 5 seconds. Brand by second 3. CTA at 15s and 30s. 15-60 second optimal length.
• Bumper (6s): Single message. Brand recall focus. Use for retargeting and frequency building.
• Shorts Ads: Vertical 9:16 format. Native-feeling creative. Under 60 seconds.
• Targeting: Custom intent (search terms), in-market audiences, competitor channel placements, topic targeting, remarketing.

• Feed: Optimize titles (brand + product + attributes), descriptions, GTINs, product categories, images (white background, 800x800+), sale prices, availability.
• Structure: Segment by product margin, performance tier, or category. Separate brand vs non-brand if using Standard Shopping alongside PMax.
• Bidding: tROAS for established products with conversion history. Maximize conversion value for new products gaining data.

• CBO (Campaign Budget Optimization): Default for 3+ ad sets. Let Meta allocate budget to best performers. Set minimum spend per ad set if needed.
• ABO (Ad Set Budget): Use when testing new audiences at controlled spend. Use for retargeting with fixed daily budgets.
• Naming Convention:

  {brand}_{objective}_{audience}_{placement}_{creative-type}_{date}

  -- consistent naming enables reporting.

• Advantage+ Shopping: Feed-based. Provide 150+ catalog items. Minimal audience restrictions. Algorithmic creative and audience optimization. Best for e-commerce with conversion pixel data.
• Advantage+ App Campaigns: Mobile app installs. Automated creative combinations.
• Advantage+ Audience: Expanded targeting using audience suggestions as signals. Replaces detailed targeting expansion.

• Cold (Prospecting): Interest stacking, broad targeting with strong creative, lookalike audiences (1-3% for quality, 3-5% for scale, 5-10% for broad reach).
• Warm (Consideration): Website visitors (30/60/90/180 day windows), video viewers (25/50/75/95% watched), social engagers, email list matches.
• Hot (Conversion/Retargeting): Add-to-cart abandoners (1-7 days), checkout initiators, past purchasers for cross-sell/upsell, high-value customer lookalikes.
• Exclusions: Always exclude converters from prospecting. Exclude existing customers from acquisition campaigns unless cross-selling.

• Single Image: 1080x1080 (feed), 1080x1920 (Stories/Reels). Primary text under 125 chars for above-fold. Headline under 40 chars.
• Carousel: 3-10 cards. Tell a story or showcase products. First card must hook. Last card = CTA.
• Video: 15-30s optimal for feed. 6-15s for Reels. Hook in first 3 seconds. Captions mandatory (85% watch muted). Square 1:1 for feed, 9:16 for Stories/Reels.
• Collection: Hero image/video + product catalog below. Instant Experience landing page. Best for e-commerce.
• Reels Ads: Native-feeling, vertical, fast-paced. UGC-style outperforms polished. Trending audio when possible.

• Set up server-side event tracking alongside the pixel (redundant tracking). Deduplicate events using event_id. Track: PageView, ViewContent, AddToCart, InitiateCheckout, Purchase, Lead, CompleteRegistration. Match quality score target: 6+. Use partner integrations (Shopify, WordPress, etc.) for simplified setup. For full server-side tracking setup and consent mode details, see

  ./references/privacy-tracking.md

  .

• Sponsored Content: Single image, carousel, video, document ads in feed. Primary B2B format.
• Message Ads: Direct InMail. High open rates (40-50%). Use for high-value offers, event invites, demo requests. Frequency cap: 1 message per member per 45 days.
• Conversation Ads: Choose-your-own-path messages. Multiple CTAs. Good for qualifying leads.
• Text Ads: Right rail and top banner. Low cost, low CTR. Brand awareness filler.
• Lead Gen Forms: Pre-filled from profile data. 2-5x higher conversion than landing pages. Keep to 3-4 fields max.

• Job Title / Function / Seniority: Target decision-makers directly. Combine function + seniority for precision.
• Company Size / Industry / Name: ABM by uploading company lists. Layer company size for SMB vs enterprise campaigns.
• Skills / Groups / Interests: Reach niche professionals by skill endorsements or group membership.
• Matched Audiences: Website retargeting, contact list upload, lookalikes, ABM company lists.
• Minimum Audience Size: 50,000+ for sponsored content. 15,000+ for message ads.

• Objectives: Awareness (reach, video views), Consideration (traffic, app installs, video views), Conversion (website conversions, catalog sales, lead gen).
• Spark Ads: Boost organic posts from your account or creator accounts. Higher engagement than standard ads. Retains social proof (likes, comments, shares).
• Targeting: Interest and behavior categories, custom audiences (pixel, app, engagement), lookalikes, hashtag interaction targeting.

• Vertical 9:16 only. 15-30 seconds optimal. Native-feeling content outperforms polished ads. Hook in first 1-2 seconds. Text overlays for key messages. Trending sounds and formats. Creator/UGC style. Show the product in use within first 3 seconds. CTA in final 3 seconds.
• Creative Refresh: Every 7-14 days. Creative fatigue hits fast on TikTok. Rotate 3-5 creatives per ad group minimum.

• Product catalog integration. Shoppable video ads. Live shopping events. Creator affiliate partnerships. In-app checkout reduces friction.

• Contextual: Page content relevance. Privacy-safe. Increasing importance post-cookie.
• Behavioral: Past browsing, purchase intent signals.
• First-Party Data: CRM lists, website visitors, app users. Most valuable post-cookie.
• Geo-Targeting: DMA, zip code, radius targeting for local campaigns.

• Non-skippable 15s or 30s spots. Streaming inventory (Hulu, Peacock, Roku, Tubi, Pluto TV). Household-level targeting. Measure via brand lift studies and website visit lift. Complement with retargeting on mobile and desktop.

• Pre-bid brand safety verification (IAS, DoubleVerify, MOAT). Inclusion lists over exclusion lists. Avoid made-for-advertising (MFA) sites. Monitor placement reports weekly. Set viewability thresholds (50%+ in-view for display, 100% in-view for video).

• PAS: Pain > Agitate > Solution. Best for problem-aware audiences.
• AIDA: Attention > Interest > Desire > Action. Classic full-funnel.
• BAB: Before > After > Bridge. Transformation-focused.
• 4U: Useful, Urgent, Unique, Ultra-specific. Best for headlines.

• Lead with the benefit, not the feature. Include numbers when possible. Use power words (free, new, proven, instant, secret, limited). Test question vs statement vs command. Character limits: Google Search 30 chars, Meta 40 chars, LinkedIn 70 chars.

1. Write 1 headline + 1 primary text per angle before any testing.
2. Keep the offer, CTA, and landing page constant across angles — only vary the angle.
3. Test all angles simultaneously in Phase 1 (wide targeting, equal budget split).
4. Identify the winning angle by lowest CPA or highest CTR after 100+ conversions.
5. Move to Phase 2: iterate within the winning angle (hook variations, visuals, copy length, CTA).
6. Move to Phase 3: test winning message across formats (static, carousel, video, UGC).

./brands/{brand-slug}/campaigns/paid-ads/creative/iteration-log-{platform}-{YYYY-MM-DD}.md

• Which angle types consistently win for this brand/audience?
• Which formats win at each funnel stage?
• What hooks perform best on which platform?
• At what frequency (impressions/user) does fatigue begin?

• Phase 1 -- Concept Testing: Test 3-5 distinct angles with equal budget. Wide targeting. Winner = lowest CPA or highest CTR.
• Phase 2 -- Iteration: Take winning angle, test variations (hooks, visuals, copy length, CTA).
• Phase 3 -- Format Testing: Test winning message across formats (static, carousel, video, UGC).
• Statistical Significance: Minimum 100 conversions or 1000 clicks per variant before declaring a winner. Run tests for 7+ days to account for day-of-week variation.

• Provide multiple headlines, images, descriptions, CTAs. Platform tests combinations algorithmically. Use on Meta (Dynamic Creative), Google (Responsive Search/Display Ads). Best for high-volume campaigns. Review asset-level performance and replace underperformers.

• New brands / no data: 70% conversion, 20% consideration, 10% awareness. Build conversion data first.
• Established brands: 50% conversion, 30% consideration, 20% awareness. Balanced funnel.
• Scaling phase: 40% conversion, 30% consideration, 30% awareness. Feed the top of funnel.
• Cross-platform split: Allocate based on where the target audience lives (from SOSTAC strategy), not evenly.

• Start manual/conservative: Understand baseline CPCs and CPAs before automating.
• Graduate to automated: Once 30+ conversions per month, switch to target CPA or target ROAS.
• Set realistic targets: Use current CPA/ROAS as starting target. Improve by 10-15% increments, not 50%.
• Budget-to-bid alignment: Daily budget should support 5-10x the target CPA to exit learning phase.

• Google: Data-driven attribution (default). Last-click as fallback for low-volume.
• Meta: 7-day click, 1-day view default. Compare against 1-day click for conservative view. Use Conversion Lift studies for incrementality.
• Cross-Platform: UTM parameters on all ad URLs. GA4 as central source of truth. Compare platform-reported vs GA4 attribution. Understand each platform over-reports.

• Cold Audiences: Broad targeting with strong creative. Lookalikes from best customers. Interest stacking on Meta. Custom intent on Google.
• Warm Audiences: Site visitors (segment by pages visited, time on site). Video viewers. Social engagers. Email openers.
• Hot Audiences: Cart abandoners. Pricing page visitors. Demo requesters. Free trial users. Past purchasers for upsell.
• Exclusion Hygiene: Exclude converters from acquisition. Exclude employees. Exclude irrelevant geographies. Suppress unsubscribed contacts.

• Vertical Scaling: Increase budget 15-20% every 3-5 days on winning campaigns. Never more than 2x in a single change. Monitor CPA drift after each increase.
• Horizontal Scaling: Duplicate winning campaigns to new audiences. Test new platforms. Expand to new geographies. Launch new creative angles to same audiences.
• Creative Scaling: The number one scaling bottleneck is creative fatigue. Produce 5-10 new creatives per week when scaling aggressively. Rotate creative every 7-14 days on Meta and TikTok.

./brands/{brand-slug}/campaigns/paid-ads/

./references/deliverable-templates.md

• 8.1 Campaign Brief -- Full strategic brief covering objective, audience, structure, copy, bidding, tracking, and testing plan. Save as

  campaign-brief-{campaign-name}-{YYYY-MM-DD}.md

  .
• 8.2 Ad Copy Document -- Platform-specific ad copy with multiple angle/hook variants, character count verification, and testing priority. Save as

  ad-copy-{platform}-{campaign}-{YYYY-MM-DD}.md

  .
• 8.3 Audience Targeting Spec -- Cold/warm/hot audience definitions with targeting criteria, estimated sizes, exclusions, and lookalike strategy. Save as

  audience-spec-{platform}-{YYYY-MM-DD}.md

  .
• 8.4 Budget Allocation Plan -- Platform and funnel-stage budget splits with scaling rules and reallocation triggers. Save as

  budget-plan-{YYYY-MM}.md

  .
• 8.5 Performance Benchmarks -- Industry and funnel-stage benchmarks with optimization thresholds. Save as

  benchmarks-{platform}-{YYYY-MM-DD}.md

  . See also

  ./references/benchmarks.md

  for cross-platform industry benchmark data.
• 8.6 A/B Testing Plan -- Hypothesis, variables, control/variant setup, sample size, and decision framework. Save as

  ab-test-plan-{test-name}-{YYYY-MM-DD}.md

  .

1. Read brand context and SOSTAC (Section 0). Always.
2. Clarify scope: Which platform? Campaign type? Funnel stage? New campaign or optimization?
3. Assess current state: Check

   ./brands/{brand-slug}/campaigns/paid-ads/

   for prior work.
4. Deliver actionable output: Specific campaigns, copy, audiences, budgets -- never vague advice.
5. Save deliverables: Write all outputs to

   ./brands/{brand-slug}/campaigns/paid-ads/

   .
6. Recommend next steps: What to launch first, what to test, when to optimize.

• No conversion tracking set up -- recommend implementation before spending.
• No landing page -- recommend building one before driving traffic.
• Budget under $500/month -- recommend focusing on 1 platform only.
• Heavily regulated industry (pharma, finance, alcohol) -- flag platform-specific ad policies.
• SEO may be more cost-effective -- route to SEO specialist for organic opportunities.