Build an n8n workflow that writes your Monday marketing report
Your marketing data is scattered across four or five platforms. Pulling it together manually each week takes longer than it should, and by the time you’ve assembled the numbers, half the morning is gone. This n8n workflow fixes that: it pulls the data, writes a plain-language summary with GPT-4o, and drops the report in your inbox every Monday at 7:00 AM without you touching a thing.
Here’s exactly how to build it.
The Five Questions Your Report Must Answer
Before touching n8n, nail down what the report needs to tell you. A useful weekly marketing report answers five specific questions:
- Lead generation: How many new leads came in this week versus last week and the four-week average? Which source drove the most?
- Pipeline health: How many leads moved stages? Which stage is most stalled? How many moved from Qualified to Booked?
- Email performance: What were open and click rates across this week’s sends? Did any individual email over- or under-perform significantly?
- Ad spend and efficiency: What was total spend? What was the cost per lead? How does it compare to last week?
- Revenue: What payments came in? Which pipeline stage did they come from?
These five data points tell you whether your marketing system is working. Everything else is noise.
️ Workflow Architecture
The workflow has six stages. Three parallel data pulls feed a merge step, which feeds an AI writing step, which feeds two delivery nodes. Here’s each piece:
Step 1: Schedule Trigger
Set a Schedule Trigger node to fire every Monday at 7:00 AM. That’s the only recurring action you’ll ever take on this workflow.
Step 2: Three Parallel HTTP Request Nodes
From the trigger, three HTTP Request nodes fire simultaneously. Each calls a different API:
- GoHighLevel Opportunities API: Pulls new contacts, stage movement, and pipeline deal data from the past seven days.
- Meta Ads API: Retrieves spend, reach, clicks, and lead count from the previous week.
- GoHighLevel email stats API: Fetches open rates, click rates, and reply rates for the week’s sends.
Running these in parallel keeps the workflow fast. There’s no reason to wait for one API response before starting the next call.
Step 3: Code Node (JavaScript merge)
A short JavaScript function in a Code node merges the three API responses into one structured object with clear labels for each metric. This is the data package that gets handed to the AI.
The merge step matters. If you pass three separate, unlabeled API response blobs to the AI, you’ll get a confused summary. A clean, labeled object produces a clean, specific narrative.
Step 4: AI Agent Node
The AI Agent node uses GPT-4o to write the actual report. The system prompt is where the quality lives (more on this below).
Step 5: Gmail Node + Slack Node (parallel)
The written report goes two places at once. The Gmail node sends it to your email. The Slack node posts the key metrics to your team channel. Both run in parallel off the same AI output.
The System Prompt That Makes It Useful
This is the difference between a useful report and a wall of numbers. Use a prompt along these lines:
You are a marketing data analyst. Write a clear, specific, plain-language weekly performance summary for a service business owner. Use bullet points for key metrics. Flag anything that changed by more than 20 percent compared to last week. End with one specific recommended action based on the data. Be concise. Keep it under 300 words.
With that prompt, the report arrives in your inbox looking something like this: “This week: 14 new leads, up 22 percent from last week. Top source: Facebook, 9 leads at $18.80 cost per lead. Six proposal-stage deals have been sitting for more than seven days. Recommended action: send a follow-up email to those six contacts today.”
That takes thirty seconds to read and tells you exactly what to do next.
Customizing the Workflow
Three things are easy to adjust as your business changes:
- Metrics pulled: Add or remove API calls as your stack evolves.
- Comparison periods: Change seven days to 30 days, or add a quarter-over-quarter comparison.
- AI benchmarks: Add baseline targets directly to the system prompt. For example: “Open rates above 30 percent are good. Below 20 percent, flag as a concern.” The AI then makes qualitative judgments instead of just reporting raw numbers.
Meta Ads, Google Ads, and LinkedIn Ads all provide API access on their standard plans. If API setup feels like too much friction upfront, Make.com has pre-built module connectors for all major ad platforms as an alternative path.
Why This Beats a Dashboard
A dashboard requires you to log in, find the right view, read the numbers, and draw your own conclusions. This workflow does all of that interpretation for you and delivers the conclusion to your inbox before you’ve opened your laptop. It’s not a prettier data display. It’s a decision-support tool that runs without you.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
