# Paid Media Experiment Design

> Use for PPC work when Codex should design paid-media tests with hypotheses, holdouts, success criteria, duration, and stopping rules.

- Skill: `srednoff888-art/paid-media-experiment-design` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add srednoff888-art/paid-media-experiment-design`
- Raw SKILL.md: https://api.skillmd.com/api/skills/srednoff888-art/paid-media-experiment-design/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: srednoff888-art (https://skillmd.com/u/srednoff888-art)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/srednoff888-art/paid-media-experiment-design

---


# paid media experiment design

## Workflow

1. Confirm the business outcome, owner, source data, constraints, and success metric.
2. Inspect the project and available account or analytics evidence before proposing changes.
3. Use official documentation for current platform behavior and retain source provenance.
4. Produce an evidence-backed plan with assumptions, dependencies, and measurable validation.
5. Run only read-only diagnostics by default; gate external, paid, production, or publishing actions behind explicit approval.

## Domain Focus

design paid-media tests with hypotheses, holdouts, success criteria, duration, and stopping rules.

## Official Sources

https://yandex.com/dev/direct/doc/en/ | https://developers.google.com/google-ads/api | https://developers.facebook.com/docs/marketing-apis

## Guardrails

- Start with read-only data and recommendations.
- Require explicit approval before changing spend, bids, audiences, ads, campaigns, or external account settings.
- Preserve campaign history and report the expected impact, rollback path, and measurement limitation.