Stop 9 of 9 · Weekly
Nour's month and the capstone
The three scheduled routines that carry the month, then the graded capstone: one account from audit to review with its guardrail log.
What it is
Eight stops of parts, arranged into a month that runs whether or not Nour has a good week. Three routines carry it, and the capstone proves the whole chain on one real account.
The three routines
| When | What it does unattended |
|---|---|
| Monday 07:00 | The six-check audit on all twelve accounts, findings ranked by dollars, written to a file |
| Wednesday 06:00 | The search term and placement cleanup delegation, proposals per account, nothing applied |
| First working day 06:00 | The report pipeline over last month exports, twelve drafts plus the QA file |
Each one reads, ranks and drafts. None of them applies a bid, budget, status, negative keyword, audience or placement change, because that is the rule the whole system rests on and a schedule is exactly where people quietly let it slip.
Every Monday at 07:00, run the six-check audit across accounts C01 to C12, read only, last 28 days against the previous 28. Write one file: every finding with dollars at risk and what to check next, sorted by dollars across all accounts, plus a line per account saying which checks could not be completed. Do not apply anything and do not contact any client.
The capstone
Pick one real account. Run it end to end this month and keep the evidence. You are producing five artefacts: the ten-minute audit with its findings table, the cleanup proposal with your keep, drop and change marks, one Performance Max or Meta diagnosis with a sized test plan, the client review with a confidence line on every recommendation, and the guardrail log covering every proposal from the month.
[upload the five capstone artefacts: audit, marked-up cleanup proposal, diagnosis and test plan, client review, guardrail log] Grade this against the rubric below, out of 100, scoring each criterion separately with the evidence you used. For every point not awarded, say exactly what is missing and what would earn it. Be strict about unverified numbers and about any change applied without a logged approval.
The rubric
| Criterion | Points | Full marks means |
|---|---|---|
| Findings are verifiable | 20 | Every figure has rows, a date range and an attribution window; nothing unverified reaches the client |
| Guardrail integrity | 25 | No bid, budget, status, negative, audience or placement change without a logged human approval, and at least one refusal with a written reason |
| Proposal quality | 20 | Each proposal names the change, the dollars, the risk if wrong and the fourteen-day outcome date |
| Test is resolvable | 15 | The test plan states one variable per cell and the results each cell needs before it is read |
| Review earns the renewal | 20 | Result against goal in one sentence, the trade named, and a confidence line with a reversal trigger on every recommendation |
Eighty passes. Under eighty, the gap is almost always the same two: figures without rows, and a change applied without a logged approval.