paid-search-ads-playbook

When to Activate

Use this skill for Google Ads (Microsoft Ads analogs) planning: structure, keywords, match types, negatives, RSAs, bidding lens, and landing-page alignment. This is paid search, not organic SEO—intent is conversion and efficient spend, not rankings.

First Questions

1. Business & offer — What sells, margin, and primary conversion (lead, purchase, trial)?
2. Geo & language — Countries, cities, languages?
3. Budget & goal — Daily/monthly cap, target CPA/ROAS (or learning phase expectations)?
4. Funnel stage — Brand, non-brand, competitor, remarketing?
5. Existing assets — Live account structure (if any), feed/SAIQ, conversion tags health?
6. Landing pages — URLs per offer; mobile speed and message match?

---

Account & Campaign Structure

• Separation of intent: Brand, generic non-brand, competitor, and remarketing belong in distinct campaigns (or clear ad groups with budget control at campaign level).
• Naming: Consistent convention (geo | network | intent | offer).
• Budget: Avoid dilution—fewer strong campaigns beat many starved ones early.
• Networks: Search vs Performance Max / Display—do not mix goals blindly; call out when PMax or Display changes assumptions.

---

Keywords & Match Types

• Match types (modern Google): Exact and Phrase are precision tools; Broad requires strong negatives and monitoring.
• Themes: One tight theme per ad group (5–15 closely related keywords, not kitchen-sink lists).
• Cannibalization: Same keyword in multiple ad groups/campaigns causes auction confusion—consolidate or use negatives to sculpt.
• Expansion: Mine search terms reports (conceptually); add winners as keywords, losers as negatives.

---

Negative Keywords

• Build shared negative lists (geo exclusions, job seekers, free-only intent, irrelevant industries).
• Ad-group-level negatives to sculpt traffic between sibling groups.
• Review weekly in early phases; stale negatives hide volume.

---

Responsive Search Ads (RSAs)

• 15 headlines / 4 descriptions where possible; pin only when legally or brand-required (pinning reduces combinations).
• Coverage: At least 3 distinct value props, 2 CTAs, 1 keyword-rich headline, 1 trust/credibility.
• Policy: Avoid superlatives without proof; medical/financial sensitivity.

---

Bidding & Goals (Framing, Not Live Optimization)

• Learning: Major changes reset learning; batch edits.
• tCPA / tROAS: Need conversion volume; suggest thresholds honestly.
• Manual / eCPC: When automation lacks data.
• Portfolio: Align bid strategy with actual conversion lag and value rules.

---

Conversion Tracking & Quality

• Primary vs secondary conversions—avoid optimizing on weak proxies.
• Enhanced conversions / consent—note when relevant.
• Landing page: Message match headline ↔ ad; one clear CTA; speed and mobile UX.

---

Tools and data sources (keyword/API + page read)

• Volume & CPC signals: seo_search_volume, seo_keyword_ideas, seo_related_keywords, seo_keyword_suggestions for seeds and expansion.
• Paid estimates: seo_ad_traffic_by_keywords for impression/click/cost scenarios when comparing bid levels.
• SERP context: seo_serp_analysis when auction insight helps (ads + organic layout).
• Landing pages: web_scrape to audit copy, CTAs, and offer clarity—never guess page content without a URL.

Be explicit: tool outputs are estimates; validate in the ad platform.

---

Quality Gate (before final deliverable)

• Campaign/ad group map matches intent separation (brand vs non-brand vs competitor vs RLSA)
• Each ad group has a tight theme; no duplicate keyword conflicts across groups
• Match type choices justified; broad groups paired with negative strategy
• Negative keyword lists proposed (shared + ad-group sculpting)
• RSAs: enough unpinned assets; pins only where required
• Bidding lens matches data volume and business goal
• Conversion tracking assumptions stated; gaps flagged
• Landing page message match called out per major ad group
• Next 7-day action list for the user (what to build first in-platform)