A Marketing Performance Report Leaders Actually Read
Turn a pile of channel metrics into a report a busy leader reads in two minutes: what moved, what it means, what you're doing about it — not a wall of numbers.
A leader doesn't want your metrics. They want what changed, what it means, and what you're doing about it.
A wall of channel numbers makes a leader's eyes glaze. Store what actually matters to the business and the report becomes a short narrative: the few numbers that moved, the why behind them, and the decisions they point to.
The report nobody reads is the one that lists every metric equally. If everything's reported, nothing's important.
A performance report's job is to drive a decision, not to prove you collected data.
The numbers must be right first. A clear narrative built on a miscounted metric is a confident way to mislead — check the data before you frame it.
| Complexity | Easy |
| Tools needed | Claude Pro or above, Claude Desktop → Cowork mode, on macOS or Windows, Claude Projects |
| Time to build | ~25 min first time · ~10 min per report after that |
| Best for | Marketing manager or founder who reports to leadership — B2B companies with a small marketing function |
What this solves
Performance reports that list every metric equally, so leaders skim them and still ask how marketing is doing.
The problem
It’s reporting time and you assemble the numbers — sessions, open rates, follower counts, spend, conversions — into a deck or a doc, and send it up. Your leader skims it, can’t tell what’s good or bad, and asks “so… how are we doing?” — the exact question the report was supposed to answer. The problem isn’t missing data; it’s that the report presents every metric as equally important and leaves the interpretation to the reader, who doesn’t have time to do it. What they want is short: what moved, what it means, and what you’re going to do about it. Producing that takes real synthesis, which is why most reports stay a data dump.
The fix is turning the metrics into a decision-ready narrative, every time, without it eating your week.
Ingredients
- Claude Subscriptions — Cowork and Projects aren't available on the free plan
- Platforms & Modes — Cowork runs in the desktop app only — not on web or mobile. This recipe analyzes metrics and writes a narrative report — multi-step work that runs in Cowork
- Claude Projects — What matters to your business and how your leadership likes information stay the same. A Project stores them so every report leads with what they care about
How it works
You report on a cycle, and what matters to your business and how leadership likes information are constant. A Project stores them so every report leads with the right few numbers and speaks to your audience — not a generic dashboard.
- Open Claude Desktop and click Cowork in the mode selector across the top (Chat · Cowork · Code).
- In the left panel, find Projects and click the + button.
- Choose Start from scratch. Name the project “Marketing Reports” and let Claude set up its folder.
- You’ll know it worked when the project appears in the left panel with its own folder and an instructions field.
Your project has a folder on your computer — that’s where the rules file lives, and where each period’s metrics land.
- Ask Claude, right in the project: “Create a file in this project’s folder:
report-rules.md. Leave it empty — I’ll fill it in.” (Or create it yourself in any text editor and save it into the project folder.) - Fill it in using the description below. Each cycle, save that period’s
metrics.csvexport into the same folder (or connect your analytics — see Extra Credit). - Confirm Claude can see them: ask “List the files you can see in this project.” Both filenames should come back. If they don’t, see If It Doesn’t Work.
report-rules.md
What actually matters to the business, your targets, and how your leadership wants information — so the report ranks by importance and speaks their language.
Example: “North-star metrics: qualified leads, pipeline influenced, CAC. Vanity metrics to de-emphasize: follower count, impressions. Audience: CEO who wants three takeaways and the action, not channel-by-channel detail. Always compare to target and to last period.”
metrics.csv
Your period’s numbers across channels — exported, or pulled live if analytics is connected. Garbage in, garbage out, so make sure it’s complete.
Open your Project in Claude Cowork. Give Claude the specifics for this run, ask for the main output, then follow up for any additional pieces you want. The exact wording for each prompt — with what it’s asking for and why — is in What you actually type below.
Check three things before it goes up:
- The numbers are correct. This comes first — a clear narrative on a miscounted metric is a confident way to be wrong. Spot-check the key figures against the source before you trust the framing. Synthesis on bad data is worse than no report.
- It leads with meaning, not metrics. The top should be takeaways and what they mean, not a channel-by-channel table. If it opens with “sessions were 12,403,” it’s still a data dump. Push the interpretation to the front and the raw numbers to an appendix.
- The “why” is honest, not flattering. A report that explains every good number and hand-waves the bad ones isn’t useful. Misses should be named with a real cause and a response. If a dip is glossed over, ask for the honest read.
Before you send: does the story match what you actually know happened? If a number spiked because of a one-off you’re aware of, the report should say so, not credit it to strategy. Your context is what keeps the narrative honest.
What you actually type
Name your files explicitly the first few runs, and ask Claude to show its work on anything that matters.
“Here are this month's marketing metrics [attach CSV or pull from analytics]. Using `report-rules.md`: write a report that opens with three takeaways, focuses on our north-star metrics vs. target and last period, explains what moved and why, names the clear wins and misses, and ends with prioritized next steps. Keep it to a page. Show the underlying numbers in an appendix, not the body.”
“Draft the two-sentence Slack version for the leadership channel.”
A one-page report that opens with three takeaways, covers your north-star metrics against target and last period, explains what moved and why, names the wins and the misses honestly, and ends with prioritized next steps — raw numbers in an appendix, not the body. The follow-up adds a two-sentence Slack version for the leadership channel. If it opens with a channel-by-channel table, or every miss gets hand-waved while every win gets credit, it missed — send it back.
- It can't fix bad data. A clear narrative on a miscounted metric is a confident way to mislead — spot-checking the key figures against the source is still your job.
- It doesn't know about one-offs. A spike from a single PR hit will read as strategy unless you say so — your context is what keeps the narrative honest.
- It reports and recommends; it doesn't decide. The prioritized next steps open the conversation with leadership — they don't replace it.
If it doesn’t work
- No Cowork tab in Claude Desktop — update the app to the latest version and confirm you’re on a paid plan; Cowork isn’t on the free tier. On Windows, Cowork also needs the Virtual Machine Platform feature enabled — if the tab still won’t appear, that’s the fix.
- Claude can’t see
report-rules.mdormetrics.csv— the files aren’t in the project’s folder, or they’re in a different folder than the one the project owns. Open the project, check which folder it points to, and move the files there. Then re-run “list the files you can see.” - The report opens with a wall of channel numbers —
report-rules.mdisn’t doing the ranking. Name the north-star metrics, name the vanity metrics to de-emphasize, and restate the structure in the prompt: takeaways up front, raw numbers in the appendix. If the rules file treats every metric as equal, so will the report. - A figure in the report doesn’t match your source — the CSV was incomplete or mislabeled, and the narrative got built on it anyway. Re-export the full period, check the column headers say what the numbers actually are, and run it again. Synthesis amplifies whatever it’s built on — fix the data before you fix the framing.
Extra credit
Small additions that pay back the next time you run it.
- Analytics connector — connect your analytics so the report pulls live numbers and you skip the export-and-clean step. See the Connectors guide.
- Live dashboard — for the always-on view, build a live artifact that refreshes the metrics, and reserve the written report for the narrative and decisions.
- Scheduled cadence — set a scheduled task to draft the report on the first of each month so it’s ready for your review, not a scramble.
“Leaders don't read the report that lists everything. They read the one that tells them what to do.”
What this teaches you about Claude Cowork
The recipe is one application. The principles apply to everything you’d hand to Claude.
Synthesis is the value, not collection. Anyone can export metrics; the work leaders want is the interpretation. Storing what matters to the business lets Claude rank and explain instead of list — turning a dump into a report that drives a decision.
Right numbers first, narrative second. A compelling story on wrong data is a liability. The recipe’s discipline is checking the figures before framing them, because synthesis amplifies whatever it’s built on — including errors.
The audience shapes the report. Storing how your leadership likes information means the report meets them where they are — three takeaways for the CEO, detail in the appendix for whoever wants it. The same data, told for the reader, is what finally gets read.
Who this is for
Marketing manager or founder who reports to leadership in B2B companies with a small marketing function (5–200 employees).
The pain: Reports that take a day to build and still get skimmed
The outcome: A monthly report leadership reads in two minutes and acts on
Published July 2, 2026 · 0 views