The weekly report, on autopilot — same structure, same definitions, every Monday. Opera refreshes the numbers, appends to your Sheet and posts the summary before the team logs on.
The weekly report is the most repetitive artifact in performance marketing: the same pulls, the same columns, the same recomputations, fifty-two times a year — due at the exact moment everyone wants to discuss what the numbers mean, not assemble them.
Boundary errors (Sunday counted twice or not at all, depending on timezone), restated platform numbers (Meta's Friday looked low on Monday and right on Thursday), the wrong week's row, a formula overwritten by a hurried paste, and the silent killer: definitions applied slightly differently by whoever happened to build it that week.
"Every Monday, update the weekly report with last week's spend, installs and CAC, and post to #growth."
The report stops mutating. Same period logic, same filters, same KPI formulas, same formatting — because the run executes a stored specification instead of a person's memory of one. Week 37 is computed exactly like week 12, which is what makes the trend line mean something.
Weekly paid — Wk 24 · Spend $48.2k (+6%) · New customers 3,180 (+9%) · Blended CAC $15.16 (−2%) · ⚠ Snap CAC $18.90, above the $17.50 target
Headline, deltas, exceptions, link to the Sheet. The conversation starts at 'why is Snap up' instead of 'where's the report'.
The same workbook serves the operator and the executive if it's structured for both: the summary block and Slack message for the fifteen-second skim, the rollup and channel detail for the person who has to act. Permissions stay native Google — leadership gets the link with view access, the team edits, nobody gets a new login. The discipline that makes this work is keeping the summary generated from the rollup rather than maintained beside it: one source of numbers, two reading depths, zero opportunities for the deck version and the sheet version to disagree in a board meeting.
Holidays, a market launch, a tracking outage — some weeks aren't comparable, and the report should say so rather than pretend. The pattern: the run still executes (the data is the data), the summary carries an annotation ("Eid week — volume comparisons vs Wk 23 not meaningful"), and the annotation lands in the row's notes column so the trend line carries its own context six months later. What you never do is skip the week or hand-adjust the numbers — gaps and edits are how report histories stop being trusted.
The two-minute Monday review, codified: the period label (right week, right section), spend magnitude against expectation (catches a filter change instantly), the customer count against the platform's claimed range (catches dedup or event-name breaks), and any flag the summary raised. Four glances. Everything deeper is analysis, which now has its morning back.
The one-page contract that makes week 37 comparable to week 12:
purchase, unique users, re-engagement excludedUS_; geo = US; media sources: paid onlyTen minutes to write, signed by whoever answers for the numbers — and from then on, every dispute about the report is a dispute about a line in this block, which is a dispute you can actually end.
Once the weekly runs, the other rhythms are configuration, not new projects: a daily pulse (yesterday vs 7-day baseline, flags only) for the team channel, and the monthly close (run after restatements settle, month re-pulled whole, appended to the monthly section) for the client. One stored specification, one set of definitions, three audiences served — and none of them depend on anyone's calendar memory.
Opera is built to touch production reports and live ad accounts without breaking anything:
See this running on your own reports.A 45-minute workflow audit maps your current process and shows exactly what Opera automates — step by step.
A 45-minute teardown of how you report today: we map every step, mark what Opera automates, and send you the written spec — useful whether or not you buy.