Design paid-ad campaigns with ak:paid-ads

Use ak:paid-ads to turn an approved offer, audience, objective, platform context, and budget boundary into a paid-media strategy and creative testing package. It helps select campaign types, structure campaigns and audiences, draft copy, plan creative, define retargeting, and organize measurement and optimization decisions.

The Skill's released source assumes ad-platform context, but it does not ship a direct account-management script or prove that credentials are configured. Treat its outputs as proposals until a person approves any live operation.

Choose ak:paid-ads for strategy and creative direction

Use ak:paid-ads when

• You need a paid-media brief for Google Ads, Meta, LinkedIn, X, TikTok, or another named platform.
• You need campaign and audience structure, ad-copy variants, creative directions, retargeting stages, or a testing backlog.
• You want to diagnose supplied CPA, ROAS, CTR, conversion, spend, or funnel evidence and propose the next bounded experiment.
• You need a budget allocation proposal tied to explicit targets and evidence.

Choose another workflow when

• You need authenticated Google or Meta reporting or account changes. Use ak:ads-management and approve the exact script action.
• You need to coordinate paid media with email, organic, content, and launch assets. Use ak:campaign.
• You need a launch-level phase plan rather than a campaign-level media plan. Use ak:launch-strategy.
• You lack an approved offer, truthful claims, audience evidence, landing destination, or tracking definition. Resolve those first.

Prepare the campaign contract

Provide the platform and campaign type, primary objective, approved product or offer, landing destination, customer evidence, geography, exclusions, brand and compliance constraints, current campaign history, available creative, tracking state, budget ceiling, target metric, and review date.

If customer lists, pixels, tags, or conversion events may be involved, state the data source, consent basis, retention boundary, and who can approve upload or instrumentation. Do not include credential values or raw personal data in the prompt.

Read Runtime adapters before assuming that an integration or Hook is active.

Invoke a draft-first workflow

/ak:paid-ads "Draft a Meta lead-generation campaign for the approved operations brief. Use only supplied customer research, keep total test spend at or below $8,000, and produce copy, creative briefs, audience exclusions, and a measurement plan. Do not access or change an ad account"

/ak:paid-ads "Draft a Meta lead-generation campaign for the approved operations brief. Use only supplied customer research, keep total test spend at or below $8,000, and produce copy, creative briefs, audience exclusions, and a measurement plan. Do not access or change an ad account"

$ak:paid-ads "Draft a Meta lead-generation campaign for the approved operations brief. Use only supplied customer research, keep total test spend at or below $8,000, and produce copy, creative briefs, audience exclusions, and a measurement plan. Do not access or change an ad account"

The canonical argument shape is [platform] [campaign-type]. Add the account and authority boundary in prose even for a draft-only request.

Follow the paid-ad stages

1. Confirm goals and constraints. Define the business objective, conversion, value, budget period and ceiling, target metric, geography, claims, compliance, and non-goals.
2. Confirm the offer and audience. Ground the landing destination, offer, pain points, search or interest evidence, exclusions, previous results, and available first-party data.
3. Select the platform and structure. Compare the current platform context, then propose campaign, ad-group or ad-set, audience, placement, naming, and budget structure. Verify product names and campaign types against current official provider documentation.
4. Create the draft package. Produce copy variants, calls to action, creative directions, landing-message alignment, retargeting stages, and test hypotheses. Every proof point or urgency claim must trace to approved evidence.
5. Define measurement. Map objective-specific metrics, conversion events, UTM naming, attribution assumptions, pacing checks, reporting cadence, and the evidence needed to change or stop a test.
6. Review before operation. A person verifies the audience, data use, platform policy, rights, claims, budget, account, tracking, and activation state before handing any approved action to an operating workflow.

Keep account and spend effects separate

Separate these effects explicitly:

Review outputs and evidence

A complete result should include the campaign objective and non-goals, selected platform rationale, campaign and audience structure, exclusions, copy and creative variants, landing-message alignment, budget proposal, tracking map, test matrix, report cadence, stop conditions, and open approvals.

For any analysis, record source, date range, currency, conversion definition, attribution window, reporting delay, and data gaps. Treat proposed performance targets as planning inputs, not predictions. The Skill defines no canonical output file path, so agree on a controlled destination before writing files.

Troubleshoot safely

Know the current limits

• The packaged platform examples and thresholds are historical guidance, not a current provider contract or performance guarantee.
• This Skill does not itself establish credential availability, account access, media rights, consent, or budget authority.
• It can propose tests and optimization choices but cannot guarantee reach, acquisition cost, conversion, revenue, or return on spend.
• Cursor invocation is user-verified; provider and runtime capability remain setup-specific.

Continue with the Marketing Kit overview or compare ak:ads-management before approving live paid-media work.