What should a marketing agency report weekly?
The short answer
A useful weekly marketing report answers three things in its first five lines: what happened, why, and what's changing next week. It reports what the ad platform shows separately from what the business actually saw, since the two can disagree sharply during a tracking break or attribution shift. Reports drafted by a system and refined by a person, rather than assembled by hand each week, tend to ship on time and actually get read, because the format never drifts.
A useful weekly marketing report is not a chart dump. It answers three questions in its first five lines, before any chart: what happened, why it happened, and what’s changing about next week because of it. Everything after that exists to support those five lines, not to replace them.
Why most weekly reports fail
The default format for a weekly report is a deck of charts with a one-line caption under each. That’s a status update, not a report. It puts the burden on the reader to notice the trend, guess at the cause, and figure out what needs to happen next. A good report does that work for them and states its conclusion first… the reader shouldn’t have to reach slide twelve to learn spend is down because a campaign got disapproved.
Answer-first, then the detail
The first five lines of a weekly report should read like a short memo, not a dashboard:
- What happened. The headline number, stated plainly. Spend, revenue, CAC, whatever the account is actually managed against.
- Why. The one or two real drivers, not a list of everything that moved that week. If three things moved and one caused it, say which one.
- What we’re doing about it. A specific action, not “monitoring closely.”
Everything else in the report exists to back up those lines with evidence, not to bury them.
Separate platform numbers from business truth
A number a platform reports and the number the business actually saw aren’t always the same number, and a good weekly report says so instead of picking one and hoping. Attribution windows, view-through credit, and tracking breaks all create gaps between what an ad platform claims and what revenue actually landed. The fix is reporting both, segmented by channel, so a reader can see where they agree and where they diverge.
| What gets reported | What it answers | Where it can mislead |
|---|---|---|
| Platform-reported metrics (Google, Meta, reported separately) | What each channel’s own algorithm believes happened | Attribution windows and view-through credit can inflate or shift channel-level numbers |
| Business truth (orders and revenue from the business’s own system) | What actually landed, independent of any platform’s model | Doesn’t say which channel deserves credit, only that something happened |
| Measurement health (tracking checks, signal-integrity ratios) | Whether either number above can be trusted this week | Easy to skip in a week when nothing looks obviously broken |
Report bad weeks the same way as good weeks
The format shouldn’t change when the number is bad. A report that gets longer, hedgier, or vaguer in a down week teaches its reader to distrust the good weeks too. The discipline runs the same direction either way: state what happened, state why, state the action. A bad week reported plainly builds more trust than a good week reported with flourish.
A minimal skeleton to steal
- Headline. One sentence, the number and the direction.
- Why. One or two real drivers, named specifically.
- Action. What changes this week because of it.
- By channel. Google, Meta, and any other paid channel, each getting its own three lines above.
- Measurement note. Anything affecting trust in this week’s numbers, even “nothing to flag.”
What a weekly report isn’t for
A weekly report is not the place for a strategy pivot or a full account narrative. Big shifts in direction deserve their own conversation, with room to disagree and dig in, not a bullet buried in a Friday recap. Weekly reporting is for tracking the account against its plan and flagging anything that needs a decision before the plan changes; it’s a heartbeat, not a highlight reel.
Rise
We built our own weekly reporting as a system after our account manager left, rather than hiring a straight replacement. In its first 50 days it drafted 10 weekly reporting decks for a human to refine before anything went to a client. If your current reports read like a chart dump every week, changing the format is usually the faster fix.