How reports got bad.
Monthly reports started out as short, practical documents: what happened, what we think, what's next. Over time, they got longer. Agencies added platform screenshots, because screenshots look like work. They added vanity metrics, because vanity metrics look good. They added quarterly trend graphs, because quarterly trend graphs make the current month look less volatile. They added a methodology slide because one client once asked about methodology.
A modern monthly report is often 15 to 30 pages of mostly filler, with the useful content scattered across two or three slides buried in the middle.
You don't have time to read that. You shouldn't have to. Here's what actually matters.
Four things to look for.
One: What moved, and why
The most useful part of any monthly report is a sentence or two per metric explaining what changed and the cause. "CPA went up from $42 to $58 because we tested a new creative in Google Ads that performed worse, and we've rolled it back." That's useful. "CPA went up 38% this month" is not useful; you already know that from looking at the number.
Skip any "trend" section that doesn't explain causes. A chart that goes up means something went up. You need to know why.
Two: What isn't working
Good monthly reports include a section on things that aren't working. Experiments that failed. Campaigns that underperformed. Hypotheses that turned out wrong. Platforms that aren't returning on investment.
Most reports skip this section, because nobody wants to document failure. That's exactly why it's the most informative section to read. An agency that writes honestly about what didn't work is an agency that will also write honestly about what did.
If the report is 100% positive, the report is lying by omission.
Three: What's changing next month
Before you finish the meeting, you should know what the agency is doing differently in the next 30 days. Not "continued optimization" or "testing new creative." Specific things. New campaign types. Different audiences. Shifted budget between platforms. Paused campaigns. New conversion events.
"We're going to rebalance the Meta campaign structure to consolidate two underperforming ad sets and redirect that $4,000 of budget toward the winning audience. We'll report on the result in four weeks." That's a plan.
"Continued testing and optimization across platforms" is not a plan. It's a placeholder.
Four: What the agency needs from you
A good report ends with a short list of open questions or requests. Access to something. A decision about something. Input on something. Creative assets they're waiting on. Data they can't get without your help.
If the report never has asks, either the agency is working in isolation (bad) or afraid to surface friction (also bad). A working partnership has friction. Small, resolvable friction should show up as asks.
What to skip.
- Platform screenshots. Unless there's a specific reason to look at the platform UI, the screenshot is filler. The number matters; the screen it came from doesn't.
- Quarter-over-quarter trend charts that don't cite causes. Noise.
- ROAS without context. ROAS has to be read against something: target ROAS, last quarter, competitor benchmarks, payback period. ROAS by itself, on a slide by itself, tells you almost nothing.
- Glossary sections. If the report needs a glossary, the report needs a rewrite.
- Anything labeled "insights" that isn't actually a decision or recommendation. "Insight" is a word agencies use when they have nothing to say. A real insight points at a decision.
The test.
Read the report. Ask yourself: "What decisions am I making this month based on what this says?"
If the answer is clear (we're moving budget from A to B, we're pausing this campaign, we're reviewing the creative brief) the report did its job.
If the answer is "nothing specific, things are moving along, I guess," the report failed, even if the numbers are fine.
Reports that describe motion without driving decisions waste everyone's time. The agency spent hours writing it. You spent minutes skimming it. Nothing changed as a result. That's the worst outcome.
What a good report looks like.
Two to four pages. Sometimes one page, if the month was quiet.
A section on what moved and why. A section on what's not working. A section on what's changing next. A short list of asks. A data appendix for anyone who wants to look at numbers, but kept separate from the narrative.
Everything written, not just charted. Charts get skimmed. Sentences get read. The highest-value reporting is the review, not the dashboard.
If your current agency isn't delivering something that looks like this, you're not getting the most valuable part of the engagement. The dashboard is the working surface; the review is the artifact.
Work with a senior practitioner.
Pacific Northwest Digital Marketing runs paid media for small and mid-sized businesses. Every engagement is run by a senior practitioner from first call through monthly reporting.
Start the conversation →