paid-ads

Overview

What the paid-ads skill does

The paid-ads skill turns your agent into an expert performance marketer focused on planning, launching, and optimizing paid advertising campaigns across major platforms:

• Google Ads (search, display, YouTube)
• Meta Ads (Facebook/Instagram)
• LinkedIn Ads
• Twitter/X Ads
• With references that also touch TikTok and other channels

It is designed to help when you’re working on:

• PPC and paid media strategy
• Campaign goals (CPA, ROAS, lead cost, revenue targets)
• Audience and keyword targeting
• Budget allocation and scaling plans
• Ongoing optimization and troubleshooting

The skill uses structured checklists, audience frameworks, and copy templates stored in the repository so your agent can walk you through proven, repeatable workflows instead of giving generic ad advice.

Who the paid-ads skill is for

Use the paid-ads skill if you are:

• A performance marketer or paid media specialist managing Google, Meta, LinkedIn, or Twitter/X campaigns
• A founder, marketer, or agency setting up paid acquisition channels for B2B SaaS, e‑commerce, lead gen, or apps
• A content or growth marketer who needs a systematic way to turn product positioning into testable ad campaigns

It is especially useful if you:

• Need to decide where to advertise (which platforms, what mix)
• Want guidance on campaign structure, naming, and testing roadmaps
• Care about measurable outcomes like CPA, ROAS, cost per lead, or demo requests

If you primarily want bulk creative generation (lots of ad variations) or deep landing page optimization, this skill will help with strategy but delegates heavy lifting to other skills.

What problems it solves

The paid-ads skill is built to solve common paid media problems, including:

• Unclear strategy – not knowing which platforms to use or how to structure campaigns
• Weak targeting – broad or unfocused audiences that waste budget
• Underperforming metrics – high CPC, high CPA, low ROAS without a clear diagnosis plan
• Messy setup – missing tracking, broken pixels, or incomplete platform configuration
• Ad copy struggles – difficulty turning product benefits into compelling ads

Repository references provide:

• A Before Starting checklist to gather product and campaign context
• Structured questions around campaign goals, product & offer, and audience
• Detailed audience targeting guidance per platform
• Ad copy templates for high-converting, testable creatives
• Platform setup checklists to reduce tracking and configuration mistakes

When paid-ads is or isn’t a good fit

Use the paid-ads skill when:

• You mention terms like “PPC,” “paid media,” “Google Ads,” “Facebook ads,” “LinkedIn ads,” “Twitter ads,” “ROAS,” “CPA,” “ad spend,” “ad budget,” “retargeting,” or “audience targeting.”
• You’re planning or optimizing paid campaigns for acquisition (leads, demos, purchases, installs).
• You want a structured workflow to go from product context → platform choice → campaign and audience design → optimization.

Consider complementary skills when:

• You want bulk ad creative generation or rapid creative iteration – use the ad-creative skill.
• You need deep landing page conversion rate optimization – use the page-cro skill.

If you only need organic content marketing (blog posts, SEO, social content without spend), a general content or SEO skill may be more appropriate than paid-ads.

How to Use

Installation and setup

1. Install the paid-ads skill

Install the skill directly from the repository:

npx skills add https://github.com/coreyhaines31/marketingskills --skill paid-ads

This pulls in the paid-ads skill definition plus its evals and references so your agent can follow the documented workflows.

2. Key files to review

After installation, open these files to understand how the skill thinks and what resources it uses:

• SKILL.md – Core description of the paid-ads skill, including its role as an expert performance marketer and the Before Starting context checklist (campaign goals, product & offer, audience).
• evals/evals.json – Example prompts and expected behaviors that show how the skill should handle strategy, metrics evaluation, and planning.
• references/ad-copy-templates.md – Ad copy formulas and templates for search and social ads, including CTAs and platform-specific tips.
• references/audience-targeting.md – Audience strategies for Google Ads, Meta, LinkedIn, Twitter/X, TikTok, including audience types and sizing guidelines.
• references/platform-setup-checklists.md – Step-by-step setup lists for Google Ads, Meta, LinkedIn, Twitter/X, and TikTok, plus a universal pre-launch checklist.

Use these references as the backbone for your agent’s guidance. Instead of inventing new frameworks, the agent should lean on these documents.

3. Connect to your own workflow

You do not have to copy the repository structure exactly, but you’ll get the best results if you:

• Store your own product marketing context in a file such as .agents/product-marketing-context.md (or .claude/product-marketing-context.md in older setups), as referenced in SKILL.md.
• Keep your ad accounts, analytics, and tracking accessible so you can quickly apply the setup checklists.
• Use the skill’s language (goals, offers, audiences) in your prompts so the agent can follow the expected workflow.

Running paid-ads workflows

1. Start with campaign context

The skill expects you to define three core areas before diving into tactics. You can do this in one prompt, like:

“Using the paid-ads skill, help me plan campaigns. Goal: generate demo requests with a $150 target CPA and $15k/month budget. Product: B2B SaaS for HR teams, $99/month per seat, landing page at https://example.com/demo. Audience: HR directors and people ops leads in US companies with 50–500 employees.”

Behind the scenes, the agent should align with the Before Starting section in SKILL.md, covering:

• 1. Campaign Goals – objective, target CPA/ROAS, budget, constraints
• 2. Product & Offer – what you’re promoting, landing page, value proposition
• 3. Audience – who you’re targeting and why

The skill is designed to either read an existing product-marketing context file or ask follow-up questions only where information is missing.

2. Decide platforms and structure

Once context is clear, use paid-ads to choose platforms and campaign architecture. For example:

“Using paid-ads, recommend which platforms to use for this B2B HR SaaS demo campaign, how to structure campaigns and ad groups, and how to allocate a $15k/month budget.”

The expected flow (in line with evals/evals.json) is that the agent will:

• Apply a platform selection framework (e.g., LinkedIn for B2B job title targeting, Google Ads for high-intent search, Meta and Twitter/X for retargeting and broader reach).
• Outline campaign and ad group structure, including naming conventions.
• Propose a budget split across platforms and campaigns.
• Define success metrics (CPA, ROAS, cost per lead, demo volume) and a simple attribution approach.

3. Design audiences using the references

Use the audience targeting reference to refine who you show ads to. For example:

“Using paid-ads and references/audience-targeting.md, design audiences for Google Search, Meta, and LinkedIn for this campaign. Include match types, audience sizes, and any exclusions.”

The agent can then pull from the reference to:

• Suggest keyword match types and audience layering for Google Ads
• Define core, custom, and lookalike audiences for Meta Ads
• Recommend job-title and company-based targeting for LinkedIn
• Add exclusions (e.g., existing customers, irrelevant industries, recent applicants) to protect budget

4. Create and test ad copy

Use the ad copy templates reference to turn your offer into testable creatives. Example prompt:

“Using paid-ads and references/ad-copy-templates.md, draft 3 Google Search ads and 3 Meta Ads variations for this offer, using PAS and Before-After-Bridge formulas. Include headlines and descriptions.”

The agent should:

• Choose appropriate copy formulas (PAS, BAB, Social Proof, Feature-Benefit)
• Propose headline and primary text tailored to search vs social
• Suggest CTAs that match your funnel stage
• Recommend a copy testing order so you’re not testing everything at once

5. Run setup and pre-launch checklists

Before turning campaigns live, have the agent walk through platform setup and pre-launch checks:

“Using paid-ads and references/platform-setup-checklists.md, audit my Google Ads and Meta setup. I want to confirm tracking, audiences, and launch readiness.”

The agent can reference the checklists to verify:

• Account foundations (billing, access, time zone, currency)
• Conversion tracking and tags/pixels
• Analytics integrations (e.g., GA4 + Google Ads linking)
• Remarketing audiences and exclusions
• Ad extensions, brand protection, and compliance items

6. Evaluate performance and optimization

Once campaigns are running, you can ask paid-ads to interpret metrics and propose optimizations, using a prompt similar to the evals examples:

“Using paid-ads, evaluate this performance: Google Ads CPC is $12 and cost per lead is $180. We’re getting ~80 leads/month from a $15k budget. Is this good, and what should we adjust?”

The skill is expected to:

• Put metrics in context (conversion rates, LTV, pipeline quality)
• Identify whether CPC, conversion rate, or targeting is the main issue
• Recommend bidding, budget, and audience changes
• Suggest testing priorities (creative, landing page, offer) in a structured way

Adapting the skill to your stack

You can combine paid-ads with your own tools and repositories by:

• Adding account-specific notes or constraints (e.g., compliance, regulated industries) as additional reference files.
• Storing winning audiences and ads in your own references folder so the agent can reuse patterns.
• Integrating with reporting workflows, where the agent reads performance exports and applies the same optimization frameworks defined in the skill.

FAQ

What is the paid-ads skill used for?

The paid-ads skill is used to plan, build, and optimize paid advertising campaigns across platforms like Google Ads, Meta, LinkedIn, and Twitter/X. It focuses on strategy, targeting, budgeting, bidding, and performance evaluation rather than bulk creative generation or landing page CRO.

How do I install the paid-ads skill?

Install it with:

npx skills add https://github.com/coreyhaines31/marketingskills --skill paid-ads

Then review SKILL.md, evals/evals.json, and the references/ files so your agent behavior aligns with the documented workflows.

Which ad platforms does paid-ads support?

The skill is designed around major ad platforms covered in the repository references:

• Google Ads (search, display, YouTube)
• Meta Ads (Facebook/Instagram)
• LinkedIn Ads
• Twitter/X Ads
• TikTok Ads (via audience and setup references)

Guidance is platform-specific where appropriate (for example, keyword match types for Google Ads, job-based targeting for LinkedIn, and audience types for Meta and TikTok).

Does paid-ads write ad copy for me?

Yes, but in a structured way. Using references/ad-copy-templates.md, the skill can:

• Apply proven formulas like PAS and Before-After-Bridge
• Draft headlines, descriptions, and primary text for search and social
• Suggest CTAs and a copy testing approach

If you need high-volume, automated creative generation or hundreds of variations, pair paid-ads with the ad-creative skill.

Can paid-ads optimize my landing pages too?

The paid-ads skill is aware that landing page quality affects ad performance, and it will consider landing page fit when discussing CPA or ROAS. However, in-depth landing page testing and conversion rate optimization are intentionally delegated to the page-cro skill. Use both if you want a full funnel approach.

How does paid-ads handle B2B vs B2C campaigns?

The skill is designed to adapt based on your context. For B2B (like the B2B HR SaaS example in evals/evals.json), it will emphasize platforms such as LinkedIn and Google Search with job-based and intent-based targeting. For B2C or e‑commerce, it can lean more on Meta, TikTok, YouTube, and broader audience strategies described in the references.

Do I need a product marketing context file for paid-ads to work?

You can use paid-ads without one, but SKILL.md explicitly advises checking for .agents/product-marketing-context.md (or .claude/product-marketing-context.md in older setups) first. Keeping such a file helps the agent:

• Avoid repeating basic questions
• Stay consistent with your positioning and messaging
• Move faster into campaigns, audiences, and testing

If the file is missing, the skill will ask you for the core context directly (goals, product, audience).

How does paid-ads help with reporting and optimization cycles?

Using the structures implied by evals/evals.json and the references, paid-ads can:

• Interpret performance metrics (CPC, CTR, conversion rate, CPA, ROAS)
• Diagnose whether issues stem from traffic quality, creative, landing page, or offer
• Propose prioritized optimization steps instead of random tweaks
• Recommend scaling plans when performance is strong

You can paste in performance summaries or exports, and ask the skill to analyze them using the same frameworks it uses for planning.

When is paid-ads not the right tool?

The paid-ads skill is not ideal when:

• You’re focused solely on organic marketing (SEO, organic social, content marketing without ad spend).
• You need a dedicated analytics or BI solution rather than strategic guidance.
• You want only design help (images, video editing); paid-ads focuses on strategy, structure, and copy, not creative asset production.

In those cases, pair it with more specialized skills or tools, and reserve paid-ads for the paid media strategy and optimization layer.

Where can I see more about how the skill should behave?

Open evals/evals.json in the repository. It contains example prompts and detailed expected outputs that demonstrate how the paid-ads skill should:

• Check for product marketing context
• Apply platform selection logic
• Define campaign structures and naming
• Choose audience strategies per platform
• Allocate budgets and define success metrics
• Propose a starting structure and scaling plan

Use those examples as templates for your own prompts and workflows.