SuperAPI AI Directory

paid-search-advertising

Marketing & growth

View repository

Updated 10-16 03:40

About this skill

Paid Search Advertising

Plans and specifies Google Ads programmes that harvest existing demand: intent-led keyword architecture, responsive search ads, Performance Max and Demand Gen used for the right job, clean conversion tracking, and landing pages that match the query. Output is a build specification and optimisation plan; changes to live accounts require explicit client authority.

Use When

  • We need a Google search advertising plan, account structure, keyword map or build specification.
  • Audit our existing Google Ads account read-only and give a fix list ranked by money at risk.
  • Write RSA headlines and descriptions, sitelinks, callouts or search-theme inputs for a campaign.
  • Should this objective run on Search, Performance Max or Demand Gen?
  • Test propositions or keywords with search ads before investing in SEO or a product build.

Do Not Use When

  • playbook-paid-social-advertising for Meta, TikTok or LinkedIn advertising.
  • seo-geo-optimisation or ai-generative-search-optimisation for organic search or AI-search visibility.
  • ad-to-site-journey-handoff for the landing-page brief handed to website-skills.
  • Stop at an approval-ready specification when asked to change bids, budgets or campaigns in a live account without written authority.

Required Inputs

Artefact Source/provider Required? If absent
Objective, conversion definition and value per conversion Client owner, advertising-strategy-and-budget Yes Stop; a search plan without a conversion definition optimises for clicks.
Budget, allowable CPA or target return and gross margin Client finance owner; advertising-attribution-and-measurement Yes Compute the break-even line from margin if supplied; otherwise mark the target not assessed.
Account access (read-only) or exports: search terms, keywords, conversions Client Google Ads account Conditional (audits) Produce a plan from research only and mark audit checks not assessed.
Landing pages and tracking status Client website, tag manager, website-skills Yes Plan a tracking brief and landing-page brief before launch.
Keyword research data Google Ads Keyword Planner or an approved tool, run on the day Yes Record the tool, date and locale; never invent search volumes.
Market and consent context (EEA/UK/CH traffic?) Client, media plan Conditional Treat consent requirements as unassessed and block EEA-targeted launch.

Workflow

  1. Confirm the objective, conversion action, value, margin and geography; stop if the conversion cannot be tracked or defined.
  2. Map intent: list the buyer's search jobs (problem, solution, brand, competitor, local, price) and build keyword themes from research run on the day (search build specification).
  3. Choose campaign types by job: Search for known intent; Performance Max for inventory-wide conversion with feeds or strong assets; Demand Gen for visual demand generation (register AD-05, checked 2026-09-23).
  4. Draft the account structure, negatives, budgets and bidding approach; phrase bidding and match-type behaviour as checks against the live help centre.
  5. Write RSA assets and extensions with ad-copy-and-hook-lab; check message match against each landing page through ad-to-site-journey-handoff.
  6. Specify tracking: conversion actions, values, offline import if relevant, UTMs, and consent handling (Consent Mode v2 for EEA/UK/CH users, register CW-10). Withhold launch if tracking fails QA.
  7. Set the test and optimisation cadence (search optimisation and audit); define kill and scale lines from the break-even CPA.
  8. Deliver the build specification for approval. After approval and launch by the authorised operator, review search terms, conversions and quality diagnostics on cadence; correct and rerun the checks after each change.

Outputs

Artefact Consumer Acceptance condition
Search build specification Authorised account operator, client approver Campaigns, ad groups or asset groups, keywords, negatives, budgets, bidding approach, assets, URLs and tracking listed; volatile behaviour flagged as checks
Keyword and intent map Strategist, website team Themes by intent with source tool and date; no invented volumes
Tracking and consent brief Web developer or tag owner Conversion actions, values, test method and consent handling defined
Audit report (existing accounts) Client owner Findings ranked by money at risk, each with evidence and a fix
Optimisation plan Operator and reporting owner Cadence, lines in the sand, kill and scale rules

Evidence Produced

Evidence Format Acceptance condition
Research log Tool, locale, date, queries Reproducible
Tracking QA record Test conversion screenshots or exports, date Each conversion action fires once per real action
Decision record Table of choices with rationale Every campaign-type and bidding choice has a reason and a check owner

Capability and Permission Boundaries

Read and search only; analysis is read-only. Publishing, spend, live account changes, outreach and personal-data processing need explicit, action-specific client authority. Creating or editing campaigns, budgets, bids, conversion settings or tags also needs a named operator, and uploading customer lists requires a lawful basis and consent under the client's data-protection obligations (PL-01, PL-02 for Uganda and Kenya).

Degraded Mode

Without account access, current help-centre checks or keyword tools, return the narrowest qualified result and mark the affected checks not assessed. Intent themes without volumes and structure without bids can still be delivered, with all platform-behaviour lines marked; never state a CPC, CTR, conversion rate or Quality Score benchmark from memory.

Decision Rules

Condition Action Failure or risk avoided
No working conversion tracking Build tracking first; launch only for measured tests Optimising for clicks and wasting budget
Brand terms mixed with generic terms Separate brand and non-brand campaigns Brand clicks masking poor generic performance
Performance Max proposed without feeds, assets or conversion volume Start with Search on core intent; add PMax once signals exist Opaque spend with no learning
The client sells physical products and wants Shopping, free listings or Performance Max with a feed Build and audit the Merchant Center feed with shopping and feeds: required attributes, GTIN rules and diagnostics; target Uganda, Kenya or Tanzania only (Rwanda is not a supported Shopping country) (register MERCHANT-CENTER-PRODUCT-DATA) Disapproved products, invented GTINs and campaigns aimed at an unsupported country
Standalone Display campaign requested Plan new display work in Demand Gen (AD-05, Display moving into Demand Gen from June 2026) Building on a format being retired
Search terms show irrelevant queries Add negatives and tighten themes before raising budget Paying for the wrong intent
CPA above break-even for the agreed window Diagnose query, ad, landing page and offer in that order; pause if unresolved Scaling a losing campaign
EEA/UK/CH traffic in scope and consent not implemented Block launch to those users until Consent Mode v2 and a CMP are live (CW-10) Audience exclusion and compliance risk

Quality Standards

  • Every keyword theme maps to one landing page whose headline echoes the query.
  • CPA = CPC ÷ conversion rate; set the allowable CPA from margin and lifetime value before launch.
  • Separate demand harvesting (search) from demand generation (Demand Gen, video, social) in targets and reports.
  • Record the date of every platform check; re-check before each build.
  • Apply anti-ai-slop to all ad text; apply the ethics filter to offers.

Anti-Patterns

  • Judging campaigns on clicks. Fix: judge on qualified conversions and CPA against break-even.
  • One ad group for every service. Fix: themes by intent with matching ads and pages.
  • Sending all ads to the home page. Fix: one landing page per theme via the journey handoff.
  • Broad targeting with no negatives. Fix: weekly search-terms review and a shared negative list.
  • Copying a US benchmark for Uganda. Fix: derive the client's own baseline in the first 2–4 weeks.
  • Changing bids and ads in the same week. Fix: one variable per test window.
  • Automated bidding on almost no conversions. Fix: check the current minimum-data guidance and use manual or simpler strategies until signals exist.

References