Skip to content

paid-ads

Create, optimize, and scale paid advertising campaigns across Google Ads, Meta, LinkedIn, and other platforms. Covers campaign strategy, ad creation, audience targeting, and optimization.

  • Area: Marketing
  • Category: marketing ads
  • Level: beginner
  • Input: conversation
  • Output: document
  • Setup: none
  • Roles: Head of Marketing

Plans paid advertising campaigns, writes ad copy, defines audience segments, structures tests, analyzes performance metrics, and recommends bidding and optimization…

How you ask

  • “Plan a Google Ads campaign for our SaaS product launch”
  • “Write ad copy for our Facebook retargeting campaign”
  • “Analyze our ad performance and recommend optimizations”
  • “Structure an A/B test for our LinkedIn ad creative”
  • “Design audience segments for our Meta ads campaign”

Inputs

  • profile — Business model and product type; carries the regions[] field that gates the regional-context check.
  • icp — Target audience the campaign is built around.
  • marketing-context — Positioning and channel mix the platform choice has to stay consistent with.
  • region-context — Regional business environment for the target market — Three-Layer Context layer 2; produced by regional-context.
  • objective — Primary campaign goal. It decides which platforms are viable, which bid strategy is offered, and which metric the plan is graded on. (awareness | traffic | leads | sales | app-installs)
  • budget — Spend available per month or week, used to size the 70/30 proven-versus-testing split and the number of ad sets.
  • target-cpa — Target CPA or ROAS the structure must be buildable against. Absent, the plan names the assumption it used instead.
  • request — The planning question to answer, including platforms, offers, or budgets already ruled in or out.
  • scope — company, or a client-id under {company.clients}/ — a string with one reserved value, not a closed set
  • profile — Product and business model the copy is describing.
  • icp — Audience whose problem the hook has to name in their own words.
  • marketing-context — Positioning and differentiation the copy must not contradict.
  • region-context — Regional advertising rules and claim restrictions that constrain what the copy may say.
  • platform — Ad platform the copy is written for. It fixes the character limits, asset counts, and native tone — Google RSA output follows references/rsa-output-spec.md. (google-ads | meta | linkedin | twitter-x | tiktok)
  • offer — What the ad promotes — product, free trial, lead magnet, or demo — and why it is compelling.
  • landing-page — Destination URL the copy must stay consistent with, so the ad promise and the post-click page match.
  • request — The creative brief — variant count, framework preference, and any angle already ruled out.
  • profile — Business model and product type that set which targeting signals are meaningful.
  • icp — Ideal customer profile the segment definitions are derived from.
  • marketing-context — Channels and competitor set used to size audiences and choose exclusions.
  • region-context — Geographic and regulatory limits on targeting in the market being advertised into.
  • platform — Ad platform the segments are built in. It fixes which targeting primitives exist — keywords, interests, job titles — and the minimum viable audience size. (google-ads | meta | linkedin | twitter-x | tiktok)
  • customer-data — Existing customer export used as the lookalike seed and the suppression list. Absent, the route builds interest or keyword audiences instead and says so.
  • request — The targeting problem to solve — new segments, exclusion cleanup, or retargeting window design.
  • performance-data — Campaign metrics — impressions, clicks, conversions, spend, CPA, CTR, frequency
  • performance-export — Platform reporting export the metrics are quoted from, so every figure carries a named source, attribution window, and date range per the evidence-discipline constraint.
  • target-cpa — Target CPA or ROAS goal
  • aov — Average order value for profitability check

Outputs

  • Headline, description, and CTA variants per platform, each labelled with the framework and angle it tests. scope is always required, so this route never falls back to the scratch area (output.scratch).
  • Per-campaign Red/Yellow/Green grade with its kill, hold, or scale action, the fatigue differential result, funnel diagnosis, and the profitability check. scope is always required, so this route never falls back to the scratch area (output.scratch).
  • Platform selection, account and campaign structure, naming convention, budget allocation, test plan, and the measurement setup the campaign is graded on. scope is always required, so this route never falls back to the scratch area (output.scratch).
  • Segment definitions with their build criteria, lookalike seed and rationale, retargeting windows by funnel stage, and the exclusion list. scope is always required, so this route never falls back to the scratch area (output.scratch).

Seen in these workflows