agentleFS
kelpi-aikelpi-ai/meta-ads-skills/skills/performance-analyst/SKILL.md

performance-analyst

Skill3 starsChanged 5 months agoMITView on GitHub

Read the account like an analyst who is not trying to keep the account: real numbers by angle, platform-ROAS caveats, and the three decisions the data actually supports. Use weekly, or before any scaling or kill decision.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before any big decision: scale, kill, or rebuild.

    line 13
  2. Distinctive:

    Never present platform ROAS as business truth.

    line 31

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Performance Analyst

Doctrine

Platform dashboards are a signal, not the truth: Meta grades its own homework, attribution flatters, and "great ROAS" can hide zero incremental revenue. The analyst's job is to say what the data supports, say what it does not, and refuse to fill gaps with optimism. Vanity metrics (reach, impressions, clicks) answer "did people see it"; the business question is "did the loop close": what did a customer cost, what did they pay you back, how fast.

When to use

  • Weekly, same day, 15 minutes.
  • Before any big decision: scale, kill, or rebuild.
  • Needs a Meta Ads MCP, read access.

Run it

Weekly read, READ-ONLY:

1. Last 30 days by ad, sorted by spend: spend, results, cost per result, CTR, CPM, frequency. Roll it up by ANGLE, not just by ad.
2. My three best and three worst spend allocations, in plain sentences ("$X went to Y and bought Z").
3. Then the honest caveats, every time:
   - which numbers depend on Meta's attribution (and would look different in my store/CRM data),
   - where the sample is too small to conclude anything,
   - what I would need to track to know the truth (blended cost per customer: total spend / total new customers, from MY records, not Meta's).
4. End with the at-most-3 decisions this data actually supports, and the decisions it does NOT support yet.

Guardrails

  • Never present platform ROAS as business truth. Pair it with the blended number or mark it unverified.
  • "The sample is too small" is a finding, not a failure. Say it whenever it is true.
  • Roll up by angle: individual ad noise hides the message-level pattern that actually matters.

Good output looks like

A page a busy owner reads in two minutes: what the money did, what we actually know vs what Meta claims, and 1-3 supported decisions.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.