---
name: "Canonical NotFair workflow"
repo: https://github.com/nowork-studio/notfair-plugin
source: https://github.com/nowork-studio/notfair-plugin/blob/HEAD/skills/paid-ads-amazon/SKILL.md
license: "MIT"
install: "npx skills add nowork-studio/notfair-plugin --skill paid-ads-amazon"
page: https://100xlab.com/resources/agent-skills/nowork-studio-notfair-plugin--paid-ads-amazon--xnps0k
---

# Canonical NotFair workflow

Read [`../../paid-ads/paid-ads-amazon/SKILL.md`](../../paid-ads/paid-ads-amazon/SKILL.md) completely, then follow it as the active workflow. Resolve every relative reference from that file against `../../paid-ads/paid-ads-amazon/`.

---

## Supporting file: paid-ads/paid-ads-amazon/SKILL.md

# Amazon Ads Planning

Read `../shared/operating-contract.md` and `../shared/measurement-framework.md`. This plugin does not declare a first-party NotFair Amazon Ads MCP mutation surface; use a verified connector or supplied report and deliver a reviewable operator brief.

## Start with unit economics

Record the product/ASIN, marketplace, currency, contribution margin, price, inventory constraint, and target ACoS. ACoS is spend divided by attributed ad revenue; it is only good or bad relative to margin and the user's strategic goal. Separate the advertised product from any measured cross-sell before judging performance.

Propose the narrowest learning plan: product scope, campaign intent, automatic discovery or manual term/ASIN hypothesis, budget, negative/exclusion rule, and review window. Treat search-term findings as evidence for targeted negatives or promotion into controlled targeting, not as a reason to remove broad discovery prematurely.

## Review and handoff

Report spend, attributed sales, ACoS, ROAS if useful, orders, conversion rate, search-term quality, and inventory risk for a complete comparable window. State reporting lag and attribution source. Mark changes `ready_for_review` until a verified connector or authorized Amazon Ads operator confirms the exact result.