Paid Ads Campaign Plan

Produces a complete paid campaign plan to build in your own ad account, with assumption-based projections and decision triggers. It never touches a live account.

Used by our marketing team

What It Does

Plans paid campaigns so every unit of spend has a target. From your goal, budget and offer it designs the campaign structure by funnel stage, defines audiences, proposes keywords and negative keywords, and writes ad variations to test.

It adds a conversion tracking setup, a week-by-week optimization schedule with decision triggers, and conservative, moderate and aggressive scenarios. The scenarios are assumptions, and the plan lists them.

It never logs in to an ad platform or changes a live campaign. You build the plan yourself, or hand it to whoever manages the account.

You bring the platform, objective, budget and landing page. Past campaign numbers help if you have them.

Who Uses It

Our marketing team, when starting a campaign or reviewing one that is not meeting its target.

Setup

1. Copy the skill/ folder into your agent's skills folder and name it paid-ads-specialist. For Claude Code that is .claude/skills/paid-ads-specialist/ in a project, or ~/.claude/skills/paid-ads-specialist/ for all projects. Other agents use their own skills folder.
2. Alternatively, fetch it through the Thinqmesh MCP server and let your agent install it.
3. Start a new session and ask for a campaign plan, naming the platform, goal and budget.

No scripts are needed. Never give the agent ad account access. If you want it to look at past performance, paste an exported report instead.

Example Result

Campaign 1: Search, high intent (BOFU)
Bid strategy: target CPA, set after the first weeks of data
Ad group: invoice software for small teams
Negative keywords: free, template, jobs, course

Headline options (checked against current platform limits)
1. Invoicing built for small teams
2. Send invoices in minutes
3. Start with a free demo

Optimization: days 3-5 pause ads with weak CTR; week 2 refine audiences.
Scenarios: assumptions listed under each case, not predictions.

How to Adapt

Add the platforms you use beyond the three covered. Change the optimization schedule to your review rhythm. Add your own policy checks for regulated industries.