Automating a Weekly Ads Report: What to Include and What to Skip

A weekly report that nobody acts on is a cost. Automating it makes it cheaper to produce but not more useful. The real decision is what the report should contain so that a manager reads it for two minutes and knows whether to change something.

Why most automated reports fail

They are built by exporting everything the platforms offer. The result is a long page of impressions, clicks, reach and frequency that nobody asked for, with the one number that matters buried or missing.

Two further problems appear. Each platform reports its own conversions, so adding them together counts the same sale more than once. And the report arrives without a question attached, so it describes the week instead of helping decide the next one.

A model: one question per section

Give every section a question it answers and an action it can trigger. If a section has neither, remove it.

1. 1

   Headline

   Did we make money this week? Profit or contribution, against target.

   1
2. 2

   Efficiency

   Cost per order and return on spend, from your store's orders.

   2
3. 3

   Changes

   What moved most versus last week, and the likely reason.

   3
4. 4

   Risks

   Anything abnormal: spend with no sales, a disapproved ad, broken tracking.

   4
5. 5

   Decisions

   Two or three proposed actions, each with an owner.

   5

A report that ends with decisions gets read. One that ends with a table does not.

What to include

• Business outcome from your own records. Orders and revenue from your store, not only the platform's view. See Store Sales vs Meta and Google Ads Numbers for why the two differ.
• Margin-aware efficiency. Return on spend means little without costs. Contribution Margin per Order shows how to get a figure that reflects what you keep.
• Week-on-week change with a reason. A change without an explanation prompts questions, not decisions.
• Data health. A short line saying whether tracking and data feeds looked normal.

What to skip

• Vanity totals

  Reach and impressions on their own rarely change a decision.

• Every ad's metrics

  Keep ad-level detail in the dashboard, not the weekly summary.

• Summed platform conversions

  Adding platforms double-counts the same sale.

• Daily noise

  One day's swing is rarely a signal. Compare weeks.

How to automate it

The mechanics are simple: a schedule, a data pull, a calculation, a message. n8n documents a Schedule Trigger that starts a workflow at set times, which suits a Monday-morning report. Many ad platforms also offer their own scheduled reports; check their current help pages. Check what your tools offer first; a built-in report is cheaper than a custom one.

1. Fix the definitions: what counts as an order, which date range, which time zone.
2. Pull outcome data from your store and spend from each platform.
3. Calculate the figures in one place, so every section uses the same numbers.
4. Add automated checks: if spend is above zero and orders are zero, flag it instead of printing a ratio.
5. Send to a fixed channel at a fixed time, and label the data window clearly.
6. For the commentary and proposed actions, have a person review before it goes out. Language-model drafts of the commentary are possible but can sound confident about causes they cannot know. See Safe Automation: Where a Human Must Approve.

For a wider view of what a tool like n8n handles, see What n8n Can Automate.

A practical test for the finished report: hand it to someone who did not build it and ask what they would do on Monday morning. If they cannot name an action within two minutes, a section is missing a question or a number is missing context. Revise the template, not the reader.

Trade-offs and when the advice changes

• Cadence. Weekly suits stable accounts. During a launch or a budget test you may need a daily alert while keeping the weekly report.
• Audience. An executive wants five lines. A media buyer wants detail. Do not use one report for both; link to a dashboard for depth, as in Executive Marketing Dashboard KPIs.
• Delayed outcomes. If you sell on cash on delivery, last week's orders are not final. Label them as provisional.

What the report cannot tell you

A weekly report shows what happened, not why. Attribution between platforms overlaps, late conversions change past numbers, and a week is a short window for small accounts. Do not read cause into one week's movement. Use the report to decide what to investigate, then investigate.

Next step

Write your five sections on one page and the decision each should trigger. If you want that built into a scheduled report fed by your own data, the custom reporting service covers it, and Dashboard Connecting Ads, Sales and Profitability is the right companion for the detailed view.

Sources

• Schedule Trigger (n8n documentation)