Agencies
A media report with revenue in it: the one business owners read
How to build a report that opens with money instead of reach: the five mandatory lines, what to cut, and why the order changes the conversation.
Also available in: Português · Español
Most paid media reports open wrong: page one with reach, impressions and CTR, and revenue — when it appears at all — on page six.
Whoever receives the report does not work in media. They are a business owner, a sales director, a manager accountable for budget. They have one question: did the money come back? A report that takes six pages to answer that is organized for the person who wrote it, not the person reading it.
The five lines that matter
| Line | What it answers |
|---|---|
| Spend | how much went out |
| Leads | how many contacts came in |
| Sales | how many became customers |
| Revenue | how much came in |
| Cost per sale | whether it pays |
Those five fit on one screen. Everything else is support.
The line that changes the conversation is the last one. Real cost per sale compared with the business's margin is the only metric that answers "is this worth continuing" without needing interpretation.
Order matters more than content
Build it in three layers, in this order:
Layer 1 — The result (one screen). The five lines, the period, the comparison against the previous one. Nothing else. Whoever reads only this must leave with the answer.
Layer 2 — Where the result came from. Breakdown by campaign and by source: which campaigns produced sales, which spent without producing. This is where action lives.
Layer 3 — Diagnosis. CTR, CPM, frequency, placement. This explains why layer 2 looks the way it does. It is what you use in the meeting when someone asks why, not what opens the document.
Inverting these layers is the most common and most expensive mistake: it makes it look like you are hiding the result behind activity.
What to cut
Vanity metrics. Reach and impressions in the executive summary help nobody decide. The piece on vanity metrics separates the ones that measure work from the ones that measure results.
Dashboard screenshots. A screenshot forces the reader to hunt for the number. If you need a screenshot, the number that matters was not written down.
Any metric you will not comment on. If a line exists in the report and nobody has ever mentioned it in any meeting, it is noise.
Comparing only against last month. A bad month after a great one looks like a disaster; after a terrible one it looks like victory. Show three months, not two.
The slice you have to show
The hard part, and what separates an honest report from a sales pitch: the unattributed slice.
If 30% of this month's sales carry no recorded origin, that appears in the report. Not distributed across channels, not hidden, not rounded away: one line saying 30% of sales were not attributed.
It looks like weakness and is the opposite. Showing the blind slice says you understand your own method — and buys credibility for the 70% you do claim. Whoever presents 100% attributed on 60% tracking will be contradicted by the client's finance team sooner or later.
A model that fits on one page
OCTOBER 2026 · ACME account
Investment $42,000 (Sep: $38,500)
Leads 612 (Sep: 540)
Sales 47 (Sep: 41)
Revenue $188,000 (Sep: $164,000)
Cost per sale $894 (Sep: $939)
Return 4.5x (Sep: 4.3x)
Attribution: 71% of sales with an identified source.
WHERE IT CAME FROM
Meta Ads $26,000 · 31 sales · $839 per sale
Google Ads $16,000 · 16 sales · $1,000 per sale
WHAT CHANGED
· Testimonial creative went live on the 8th and accounts for 11 of Meta's 31 sales
· Generic search campaign paused on the 14th: spent $3,100 with no sales
· Lead-to-sale rate moved from 7.6% to 7.7%
NEXT MONTH
· Scale the testimonial into the remarketing ad set
· Two new creatives in the same format
· Test branded search terms, which were left out
Three things this model does: it answers on line one, it shows coverage, and it ends with decisions instead of a summary.
The frequency mistake
A monthly report and a weekly check-in are different documents with different audiences.
The weekly is operational: what was done, what comes next, anything needing a decision. Three lines in an email will do.
The monthly is about results: the five lines, the comparison, next month's decisions.
Merging them produces a long document nobody reads in full — and worse, it pushes to month-end decisions that needed making in week two.
How to present a bad month
The honest report suffers in a bad month, and that is where trust is built or lost.
Open with the number. Hiding the result behind three pages of context costs more credibility than the result itself.
Explain the cause, not the circumstance. "The auction got more expensive in November" is a circumstance. "The auction rose 22% and we did not rotate creatives in time" is a cause — and it comes with the fix attached.
Bring the decision. A bad month with no proposed change is a report from someone who watched.
For the structure of the conversation itself, see presenting results to a client; and if the client is the one asking for metrics you consider noise, the report clients actually read covers that negotiation.
Frequently asked questions
What cannot be missing from a media report with revenue?
Spend, leads, sales, revenue and cost per sale. Five lines. Anything that does not help decide one of those five is an appendix, not the report.
What if I do not have revenue by campaign?
Show what you have and state the coverage. A report with 60% of sales attributed and that fact visible is more useful than one pretending to have 100%.
Should reach and impressions be removed?
Removed from page one. They diagnose, they do not account for results. Whoever reads the report wants to know whether the money came back, and reach does not answer that.
How often should I send it?
Monthly for results, weekly for operations. Merging the two produces a long document nobody reads in full and that delays decisions.