Aditya Sriram recently shared something that caught my attention: "10 minutes. That's all it takes to automate your weekly marketing reports forever." Then they laid out a refreshingly simple setup: "Drop in your GA4 property ID" plus your Google Ads customer ID and Meta Ads account ID, and "that's it. You're done."

I love posts like this because they cut through the usual reporting theater. Most teams do not have a reporting problem. They have a repeatability problem. The same questions come up every Monday, the same screenshots get taken, the same numbers get reconciled across platforms, and the same narrative gets rewritten by someone who would rather be improving performance.

Aditya's point is not that reporting is unimportant. It's that the mechanics of producing it can be automated, leaving humans to do what they are better at: choosing what to test next, spotting creative insights, and making decisions under uncertainty.

Key idea: automate the assembly and first draft, then spend your time on interpretation and action.

What Aditya is really proposing

In Aditya's workflow, n8n is the orchestrator. It runs on a schedule and stitches together three core steps:

Collect fresh performance data from the tools you already use.
Compute week-over-week comparisons automatically.
Use an LLM (in their example, Claude) to turn metrics into a readable narrative.

The output is not just numbers in a spreadsheet. Aditya described an end-to-end delivery system:

An HTML report with styled tables delivered to your inbox
A Slack-formatted summary posted to a channel

That last part matters. The best report is the one people actually read. Delivering it where your team already works (email and Slack) is often more valuable than building yet another dashboard.

Why weekly marketing reporting is still painfully manual

Even with excellent tooling, weekly reporting tends to stay manual for a few reasons:

The data is fragmented

Paid media performance lives in Google Ads and Meta Ads. Site outcomes live in GA4. Ecommerce revenue might be in Shopify. Attribution caveats live in everyone's heads.

When you pull metrics from multiple sources, you create two recurring tasks:

Data collection (logging in, filtering dates, exporting)
Reconciliation (making sure time windows and definitions match)

The narrative is repetitive, but still time-consuming

A standard weekly report usually includes:

What changed vs last week
What drove the change (campaigns, audiences, creatives)
What we plan to do next

The structure repeats, but writing it from scratch takes time, especially when you are sanity-checking numbers across platforms.

Dashboards do not solve the Monday morning question

Dashboards are great for exploration. Weekly reporting is different. It is a ritual and a summary. It needs a point of view and a clear "so what." Automating the report does not replace dashboards, it complements them.

The 10-minute setup, unpacked

Aditya's promise hinges on a simple configuration: IDs for GA4, Google Ads, and Meta Ads. The brilliance here is the reduction of setup friction. Less ceremony means more adoption.

Here is how the flow works in practice, expanding on Aditya's bullets.

H2: Step 1 - Trigger on a schedule

Aditya mentioned: "Every Monday at 7 AM, this n8n workflow..." That is your trigger. Weekly reporting should be boringly consistent. A cron-based trigger guarantees consistency and removes the need for someone to remember.

Practical tip: choose a time after data has settled. Some accounts see late conversions and delayed GA4 processing. Monday 7 AM can be perfect, but consider your time zone and conversion delay patterns.

H2: Step 2 - Pull live data from each platform

Aditya called out that "GoMarble MCP pulls live data from all 3 platforms". Whether you use GoMarble MCP or direct connectors, the underlying goal is the same: define a small, stable set of metrics and dimensions that answer 80 percent of weekly questions.

Common weekly metrics to standardize:

Google Ads: spend, clicks, conversions, CPA, ROAS, top campaigns or ad groups
Meta Ads: spend, impressions, clicks, purchases, CPA, ROAS, top campaigns
GA4: sessions, users, revenue, purchases, conversion rate, key landing pages

If you try to include every possible breakdown, the report becomes noisy and the automation becomes brittle. Start with a consistent core and iterate.

H2: Step 3 - Compute week-over-week comparisons automatically

Aditya emphasized: "Calculates this week vs. last week automatically" and "Week-over-week comparisons for every metric." This is the part that makes the report immediately useful.

A solid week-over-week section should include:

Absolute change (for example, spend up $2,300)
Percent change (spend up 12 percent)
A brief explanation of drivers (campaign-level deltas)

Practical tip: define the date windows explicitly in the workflow. "Last 7 days" vs "previous 7 days" is often better than calendar weeks, but it depends on your business rhythm.

H2: Step 4 - Let AI draft the narrative, not invent the numbers

Aditya wrote: "Claude analyzes performance and writes a full report." This is where teams get both excited and nervous.

The right mental model is: AI is the reporting analyst that never gets tired, but it needs guardrails.

I would structure the prompt so the model:

Summarizes what moved up or down
Attributes changes to the specific tables you provide
Flags anomalies and missing data
Suggests next steps as hypotheses, not as facts

Good prompt constraint: "Only use the numbers in the provided tables. If a driver is unclear, say so and list what extra cut you would need."

This keeps the narrative grounded and reduces hallucinations.

H2: Step 5 - Deliver in the formats people will consume

Aditya's delivery details are quietly powerful: "HTML report with styled tables" plus a "Slack-formatted version." Email is great for archiving and review. Slack is great for attention.

A practical approach:

Email: full report with tables, notes, and links to dashboards
Slack: short summary with 5 to 8 bullets, plus a link to the full report

Making the workflow actually usable in a real team

Aditya also included a helpful postscript: "Need a different report? Just change the prompt. Daily instead of weekly? Change the trigger. The workflow adapts to whatever you need." I agree, with one addition: your workflow should adapt, but your definitions should not change casually.

Here are a few suggestions to make this automation durable.

H3: Standardize metric definitions

Decide once and document:

What counts as a conversion (platform conversion vs GA4 purchase)
Which attribution windows you are using
Whether you report on click-through conversions only or include view-through (Meta)

If you do not standardize, the report will be consistent but confusing.

H3: Add basic data quality checks

Before the LLM writes anything, add checks like:

If spend is zero across all platforms, alert instead of reporting
If GA4 revenue is null, note the outage and skip conclusions
If week-over-week changes exceed a threshold, add an "anomaly" callout

H3: Keep the campaign breakdown focused

Aditya mentioned "Campaign-level Google Ads breakdowns" and "Account-level Meta Ads summaries." That split makes sense because Google Ads often needs more granular diagnosis, while Meta can be summarized at account level for weekly exec updates.

If you want a simple rule:

Include the top 5 winners and top 5 losers by spend or conversions
Include only items above a minimum spend threshold

What this means for content, not just operations

Aditya's post went viral because it is crisp, specific, and immediately actionable. "10 minutes" is a strong hook, the steps are concrete, and the CTA is simple (comment "REPORT"). It is a good reminder that the best marketing automation content:

Names a painful routine
Promises a realistic time-to-value
Shows the minimum inputs required
Describes outputs in human terms (inbox, Slack channel)

If you are building or buying reporting automation, use the same framework internally. Make the workflow easy to adopt and hard to misunderstand.

A simple next step you can take this week

If you are intrigued by Aditya Sriram's approach, try this progression:

Automate data pulls and week-over-week tables first.
Add Slack delivery with a short templated summary.
Only then add an LLM to draft narrative, with strict constraints.

You will get most of the value quickly, and you will build trust before you add generative text.

This blog post expands on a viral LinkedIn post by Aditya Sriram, Building GoMarble || AI Agent for paid media marketers; built on your Meta Ads, Google Ads, Shopify, and GA4.. View the original LinkedIn post →